4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 [xx XXX xxxx]
27 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
29 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
31 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
32 supported and enabled.
36 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
37 to the list of ciphersuites providing Perfect Forward Secrecy as
38 required by SECLEVEL >= 3.
40 *Dmitry Belyavskiy, Nicola Tuveri*
42 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting. The
43 SSL_CTRL_GET_IANA_GROUPS control code, exposed as the SSL_get0_iana_groups()
44 function-like macro, retrieves the list of supported groups sent by the peer,
45 and the function SSL_client_hello_get_extension_order() populates a caller-supplied
46 array with the list of extension types present in the ClientHello, in order of
51 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
52 to make it possible to use empty passphrase strings.
56 * RNDR and RNDRRS support in provider functions to provide
57 random number generation for Arm CPUs (aarch64).
61 * s_client and s_server apps now explicitly say when the TLS version
62 does not include the renegotiation mechanism. This avoids confusion
63 between that scenario versus when the TLS version includes secure
64 renegotiation but the peer lacks support for it.
68 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
70 *Tomasz Kantecki, Andrey Matyukov*
72 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
73 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
74 of 160 bits and above and less than 224 bits were previously accepted by
75 default but are now no longer allowed. By default TLS compression was
76 already disabled in previous OpenSSL versions. At security level 2 it cannot
81 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
86 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
87 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
88 will need to load the legacy crypto provider.
92 * The various OBJ_* functions have been made thread safe.
96 * CCM8 cipher suites in TLS have been downgraded to security level zero
97 because they use a short authentication tag which lowers their strength.
101 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
106 * Parallel dual-prime 1536/2048-bit modular exponentiation for
107 AVX512_IFMA capable processors.
109 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
114 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
115 listed here are only a brief description.
116 The migration guide contains more detailed information related to new features,
117 breaking changes, and mappings for the large list of deprecated functions.
119 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
121 ### Changes between 3.0.2 and 3.0.3
123 * Fixed a bug in the c_rehash script which was not properly sanitising shell
124 metacharacters to prevent command injection. This script is distributed by
125 some operating systems in a manner where it is automatically executed. On
126 such operating systems, an attacker could execute arbitrary commands with the
127 privileges of the script.
129 Use of the c_rehash script is considered obsolete and should be replaced
130 by the OpenSSL rehash command line tool.
135 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
136 certificate on an OCSP response. The bug caused the function in the case
137 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
138 response (meaning a successful verification) even in the case where the
139 response signing certificate fails to verify.
141 It is anticipated that most users of `OCSP_basic_verify` will not use the
142 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
143 a negative value (indicating a fatal error) in the case of a certificate
144 verification failure. The normal expected return value in this case would be
147 This issue also impacts the command line OpenSSL "ocsp" application. When
148 verifying an ocsp response with the "-no_cert_checks" option the command line
149 application will report that the verification is successful even though it
150 has in fact failed. In this case the incorrect successful response will also
151 be accompanied by error messages showing the failure and contradicting the
152 apparently successful result.
157 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
158 AAD data as the MAC key. This made the MAC key trivially predictable.
160 An attacker could exploit this issue by performing a man-in-the-middle attack
161 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
162 that the modified data would still pass the MAC integrity check.
164 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
165 endpoint will always be rejected by the recipient and the connection will
166 fail at that point. Many application protocols require data to be sent from
167 the client to the server first. Therefore, in such a case, only an OpenSSL
168 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
170 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
171 sent in both directions. In this case both clients and servers could be
172 affected, regardless of the application protocol.
174 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
175 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
176 the handshake when using this ciphersuite.
178 The confidentiality of data is not impacted by this issue, i.e. an attacker
179 cannot decrypt data that has been encrypted using this ciphersuite - they can
182 In order for this attack to work both endpoints must legitimately negotiate
183 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
184 OpenSSL 3.0, and is not available within the default provider or the default
185 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
186 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
187 following must have occurred:
189 1) OpenSSL must have been compiled with the (non-default) compile time option
190 enable-weak-ssl-ciphers
192 2) OpenSSL must have had the legacy provider explicitly loaded (either
193 through application code or via configuration)
195 3) The ciphersuite must have been explicitly added to the ciphersuite list
197 4) The libssl security level must have been set to 0 (default is 1)
199 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
201 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
202 others that both endpoints have in common
207 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
208 occuppied by the removed hash table entries.
210 This function is used when decoding certificates or keys. If a long lived
211 process periodically decodes certificates or keys its memory usage will
212 expand without bounds and the process might be terminated by the operating
213 system causing a denial of service. Also traversing the empty hash table
214 entries will take increasingly more time.
216 Typically such long lived processes might be TLS clients or TLS servers
217 configured to accept client certificate authentication.
220 *Hugo Landau, Aliaksei Levin*
222 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
223 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
224 statistics are no longer supported. For compatibility, these statistics are
225 still listed in the output but are now always reported as zero.
229 ### Changes between 3.0.1 and 3.0.2 [15 mar 2022]
231 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
232 for non-prime moduli.
234 Internally this function is used when parsing certificates that contain
235 elliptic curve public keys in compressed form or explicit elliptic curve
236 parameters with a base point encoded in compressed form.
238 It is possible to trigger the infinite loop by crafting a certificate that
239 has invalid explicit curve parameters.
241 Since certificate parsing happens prior to verification of the certificate
242 signature, any process that parses an externally supplied certificate may thus
243 be subject to a denial of service attack. The infinite loop can also be
244 reached when parsing crafted private keys as they can contain explicit
245 elliptic curve parameters.
247 Thus vulnerable situations include:
249 - TLS clients consuming server certificates
250 - TLS servers consuming client certificates
251 - Hosting providers taking certificates or private keys from customers
252 - Certificate authorities parsing certification requests from subscribers
253 - Anything else which parses ASN.1 elliptic curve parameters
255 Also any other applications that use the BN_mod_sqrt() where the attacker
256 can control the parameter values are vulnerable to this DoS issue.
261 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
262 to the list of ciphersuites providing Perfect Forward Secrecy as
263 required by SECLEVEL >= 3.
265 *Dmitry Belyavskiy, Nicola Tuveri*
267 * Made the AES constant time code for no-asm configurations
268 optional due to the resulting 95% performance degradation.
269 The AES constant time code can be enabled, for no assembly
270 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
274 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
279 * The negative return value handling of the certificate verification callback
280 was reverted. The replacement is to set the verification retry state with
281 the SSL_set_retry_verify() function.
285 ### Changes between 3.0.0 and 3.0.1 [14 dec 2021]
287 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
288 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
289 verify a certificate supplied by a server. That function may return a
290 negative return value to indicate an internal error (for example out of
291 memory). Such a negative return value is mishandled by OpenSSL and will cause
292 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
293 success and a subsequent call to SSL_get_error() to return the value
294 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
295 returned by OpenSSL if the application has previously called
296 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
297 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
298 totally unexpected and applications may not behave correctly as a result. The
299 exact behaviour will depend on the application but it could result in
300 crashes, infinite loops or other similar incorrect responses.
302 This issue is made more serious in combination with a separate bug in OpenSSL
303 3.0 that will cause X509_verify_cert() to indicate an internal error when
304 processing a certificate chain. This will occur where a certificate does not
305 include the Subject Alternative Name extension but where a Certificate
306 Authority has enforced name constraints. This issue can occur even with valid
312 * Corrected a few file name and file reference bugs in the build,
313 installation and setup scripts, which lead to installation verification
314 failures. Slightly enhanced the installation verification script.
318 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
323 * Fixed PVK encoder to properly query for the passphrase.
327 * Multiple fixes in the OSSL_HTTP API functions.
331 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
332 OSSL_PARAM_INTEGER data type and return error on negative numbers
333 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
334 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
338 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
342 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
346 * Multiple threading fixes.
350 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
354 * Allow fetching an operation from the provider that owns an unexportable key
355 as a fallback if that is still allowed by the property query.
359 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
361 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
366 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
367 S390X capability vector to zero. This simplifies testing of different code
368 paths on S390X architecture.
372 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
373 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
374 SP 800-38D". The communication will fail at this point.
378 * The EC_GROUP_clear_free() function is deprecated as there is nothing
379 confidential in EC_GROUP data.
383 * The byte order mark (BOM) character is ignored if encountered at the
384 beginning of a PEM-formatted file.
388 * Added CMS support for the Russian GOST algorithms.
392 * Due to move of the implementation of cryptographic operations
393 to the providers, validation of various operation parameters can
394 be postponed until the actual operation is executed where previously
395 it happened immediately when an operation parameter was set.
397 For example when setting an unsupported curve with
398 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
399 fail but later keygen operations with the EVP_PKEY_CTX will fail.
401 *OpenSSL team members and many third party contributors*
403 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
404 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
405 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
406 instead to retrieve these algorithms from a provider.
410 * On build targets where the multilib postfix is set in the build
411 configuration the libdir directory was changing based on whether
412 the lib directory with the multilib postfix exists on the system
413 or not. This unpredictable behavior was removed and eventual
414 multilib postfix is now always added to the default libdir. Use
415 `--libdir=lib` to override the libdir if adding the postfix is
420 * The triple DES key wrap functionality now conforms to RFC 3217 but is
421 no longer interoperable with OpenSSL 1.1.1.
425 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
426 function codes, this function can only cause problems for calling
431 * Add a configurable flag to output date formats as ISO 8601. Does not
432 change the default date format.
436 * Version of MSVC earlier than 1300 could get link warnings, which could
437 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
438 Support for this flag has been removed.
442 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
443 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
444 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
445 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
446 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
450 * The signatures of the functions to get and set options on SSL and
451 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
452 Some source code changes may be required.
456 * The public definitions of conf_method_st and conf_st have been
457 deprecated. They will be made opaque in a future release.
459 *Rich Salz and Tomáš Mráz*
461 * Client-initiated renegotiation is disabled by default. To allow it, use
462 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
463 flag, or the "ClientRenegotiation" config parameter as appropriate.
467 * Add "abspath" and "includedir" pragma's to config files, to prevent,
468 or modify relative pathname inclusion.
472 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
473 validated. Please consult the README-FIPS and
474 README-PROVIDERS files, as well as the migration guide.
476 *OpenSSL team members and many third party contributors*
478 * For the key types DH and DHX the allowed settable parameters are now different.
482 * The openssl commands that read keys, certificates, and CRLs now
483 automatically detect the PEM or DER format of the input files.
485 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
487 * Added enhanced PKCS#12 APIs which accept a library context.
491 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
495 * Added support for Kernel TLS (KTLS).
497 *Boris Pismenny, John Baldwin and Andrew Gallatin*
499 * Support for RFC 5746 secure renegotiation is now required by default for
500 SSL or TLS connections to succeed.
504 * The signature of the `copy` functional parameter of the
505 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
506 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
507 the signature of the `pub_decode` functional parameter of the
508 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
509 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
513 * The error return values from some control calls (ctrl) have changed.
517 * A public key check is now performed during EVP_PKEY_derive_set_peer().
521 * Many functions in the EVP_ namespace that are getters of values from
522 implementations or contexts were renamed to include get or get0 in their
523 names. Old names are provided as macro aliases for compatibility and
528 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
529 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
530 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
535 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
538 * The output from the command line applications may have minor
543 * The output from numerous "printing" may have minor changes.
547 * Windows thread synchronization uses read/write primitives (SRWLock) when
548 supported by the OS, otherwise CriticalSection continues to be used.
552 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
553 work on read only BIO source/sinks that do not support these functions.
554 This allows piping or redirection of a file BIO using stdin to be buffered
555 into memory. This is used internally in OSSL_DECODER_from_bio().
559 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
560 this function would return one of the values OSSL_STORE_INFO_NAME,
561 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
562 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
563 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
564 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
565 using this function should be amended to handle the changed return value.
569 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
570 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
571 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
572 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
573 contains more than one certificate identifier: This means that all
574 certificates referenced there MUST be part of the validation chain.
578 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
579 RC5, DESX and DES have been moved to the legacy provider.
583 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
584 RIPEMD-160 have been moved to the legacy provider.
588 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
593 * The deprecated functions EVP_PKEY_get0_RSA(),
594 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
595 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
596 well as the similarly named "get1" functions behave differently in
601 * A number of functions handling low-level keys or engines were deprecated
602 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
603 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
604 EVP_PKEY_get0_siphash().
608 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
609 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
610 will need to load the legacy crypto provider. This includes these PBE
611 algorithms which use this KDF:
612 - NID_pbeWithMD2AndDES_CBC
613 - NID_pbeWithMD5AndDES_CBC
614 - NID_pbeWithSHA1AndRC2_CBC
615 - NID_pbeWithMD2AndRC2_CBC
616 - NID_pbeWithMD5AndRC2_CBC
617 - NID_pbeWithSHA1AndDES_CBC
621 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
622 BIO_debug_callback() functions.
626 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
627 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
631 * The RAND_METHOD APIs have been deprecated.
635 * The SRP APIs have been deprecated.
639 * Add a compile time option to prevent the caching of provider fetched
640 algorithms. This is enabled by including the no-cached-fetch option
641 at configuration time.
645 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
646 count of PKCS12_DEFAULT_ITER.
648 *Tomáš Mráz and Sahana Prasad*
650 * The openssl speed command does not use low-level API calls anymore.
654 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
657 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
659 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
663 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
664 providers may supply their own group implementations (using either the "key
665 exchange" or the "key encapsulation" methods) which will automatically be
666 detected and used by libssl.
668 *Matt Caswell, Nicola Tuveri*
670 * The undocumented function X509_certificate_type() has been deprecated;
674 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
678 * Removed RSA padding mode for SSLv23 (which was only used for
679 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
680 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
685 * Deprecated the obsolete X9.31 RSA key generation related functions.
687 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
688 is not allowed to return a value > 1, this is no more taken as failure.
690 *Viktor Dukhovni and David von Oheimb*
692 * Deprecated the obsolete X9.31 RSA key generation related functions
693 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
694 BN_X931_generate_prime_ex().
698 * The default key generation method for the regular 2-prime RSA keys was
699 changed to the FIPS 186-4 B.3.6 method.
703 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
707 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
711 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
712 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
714 *Rich Salz, Richard Levitte, and David von Oheimb*
716 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
720 * Deprecated `OCSP_parse_url()`.
724 * Validation of SM2 keys has been separated from the validation of regular EC
729 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
730 switches: a validation failure triggers an early exit, returning a failure
731 exit status to the parent process.
735 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
736 to ignore unknown ciphers.
740 * The `-cipher-commands` and `-digest-commands` options
741 of the command line utility `list` have been deprecated.
742 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
746 * Added convenience functions for generating asymmetric key pairs:
747 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
748 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
752 * All of the low level EC_KEY functions have been deprecated.
754 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
756 * Deprecated all the libcrypto and libssl error string loading
761 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
762 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
767 * The `-crypt` option to the `passwd` command line tool has been removed.
771 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
776 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
780 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
781 EVP_PKEY_get1_tls_encodedpoint().
785 * The security callback, which can be customised by application code, supports
786 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
787 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
791 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
792 interface. Their functionality remains unchanged.
796 * Added new option for 'openssl list', '-providers', which will display the
797 list of loaded providers, their names, version and status. It optionally
798 displays their gettable parameters.
802 * Removed EVP_PKEY_set_alias_type().
806 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
807 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
811 * Changed all "STACK" functions to be macros instead of inline functions. Macro
812 parameters are still checked for type safety at compile time via helper
817 * Remove the RAND_DRBG API
819 *Paul Dale and Matthias St. Pierre*
821 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
822 as well as actual hostnames.
826 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
827 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
828 conversely, silently ignore DTLS protocol version bounds when configuring
829 TLS-based contexts. The commands can be repeated to set bounds of both
830 types. The same applies with the corresponding "min_protocol" and
831 "max_protocol" command-line switches, in case some application uses both TLS
834 SSL_CTX instances that are created for a fixed protocol version (e.g.
835 `TLSv1_server_method()`) also silently ignore version bounds. Previously
836 attempts to apply bounds to these protocol versions would result in an
837 error. Now only the "version-flexible" SSL_CTX instances are subject to
838 limits in configuration files in command-line options.
842 * Deprecated the `ENGINE` API. Engines should be replaced with providers
847 * Reworked the recorded ERR codes to make better space for system errors.
848 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
849 given code is a system error (true) or an OpenSSL error (false).
853 * Reworked the test perl framework to better allow parallel testing.
855 *Nicola Tuveri and David von Oheimb*
857 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
858 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
862 * 'Configure' has been changed to figure out the configuration target if
863 none is given on the command line. Consequently, the 'config' script is
864 now only a mere wrapper. All documentation is changed to only mention
867 *Rich Salz and Richard Levitte*
869 * Added a library context `OSSL_LIB_CTX` that applications as well as
870 other libraries can use to form a separate context within which
871 libcrypto operations are performed.
875 * Added various `_ex` functions to the OpenSSL API that support using
876 a non-default `OSSL_LIB_CTX`.
880 * Handshake now fails if Extended Master Secret extension is dropped
885 * Dropped interactive mode from the `openssl` program.
889 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
891 *David von Oheimb and Shane Lontis*
893 * Deprecated `EC_METHOD_get_field_type()`.
897 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
898 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
899 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
903 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
907 * Add CAdES-BES signature verification support, mostly derived
908 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
910 *Filipe Raimundo da Silva*
912 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
916 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
917 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
921 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
925 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
926 EC_KEY_precompute_mult().
930 * Deprecated EC_POINTs_mul().
934 * Removed FIPS_mode() and FIPS_mode_set().
938 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
942 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
943 EC_POINT_get_Jprojective_coordinates_GFp().
947 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
948 arrays to be more easily constructed via a series of utility functions.
949 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
950 the various push functions and finally convert to a passable OSSL_PARAM
951 array using OSSL_PARAM_BLD_to_param().
955 * The security strength of SHA1 and MD5 based signatures in TLS has been
960 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
961 contain a provider side internal key.
965 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
969 * Project text documents not yet having a proper file name extension
970 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
971 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
975 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
976 have been converted to Markdown with the goal to produce documents
977 which not only look pretty when viewed online in the browser, but
978 remain well readable inside a plain text editor.
980 To achieve this goal, a 'minimalistic' Markdown style has been applied
981 which avoids formatting elements that interfere too much with the
982 reading flow in the text file. For example, it
984 * avoids [ATX headings][] and uses [setext headings][] instead
985 (which works for `<h1>` and `<h2>` headings only).
986 * avoids [inline links][] and uses [reference links][] instead.
987 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
989 [ATX headings]: https://github.github.com/gfm/#atx-headings
990 [setext headings]: https://github.github.com/gfm/#setext-headings
991 [inline links]: https://github.github.com/gfm/#inline-link
992 [reference links]: https://github.github.com/gfm/#reference-link
993 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
994 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
996 *Matthias St. Pierre*
998 * The test suite is changed to preserve results of each test recipe.
999 A new directory test-runs/ with subdirectories named like the
1000 test recipes are created in the build tree for this purpose.
1004 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1005 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1006 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1008 *David von Oheimb, Martin Peylo*
1010 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1011 It supports arbitrary request and response content types, GET redirection,
1012 TLS, connections via HTTP(S) proxies, connections and exchange via
1013 user-defined BIOs (allowing implicit connections), persistent connections,
1014 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1015 The legacy OCSP-focused (and only partly documented) API
1016 is retained for backward compatibility, while most of it is deprecated.
1020 * Added `util/check-format.pl`, a tool for checking adherence to the
1021 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1022 The checks performed are incomplete and yield some false positives.
1023 Still the tool should be useful for detecting most typical glitches.
1027 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1028 If domain name resolution yields multiple IP addresses all of them are tried
1029 after `connect()` failures.
1033 * All of the low level RSA functions have been deprecated.
1037 * X509 certificates signed using SHA1 are no longer allowed at security
1042 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1043 modified to use PKEY APIs. These commands are now in maintenance mode
1044 and no new features will be added to them.
1048 * The command line utility rsautl has been deprecated.
1052 * The command line utilities genrsa and rsa have been modified to use PKEY
1053 APIs. They now write PKCS#8 keys by default. These commands are now in
1054 maintenance mode and no new features will be added to them.
1058 * All of the low level DH functions have been deprecated.
1060 *Paul Dale and Matt Caswell*
1062 * All of the low level DSA functions have been deprecated.
1066 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1067 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1071 * Deprecated low level ECDH and ECDSA functions.
1075 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1079 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1080 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1081 a new formulation to include all the things it can be used for,
1082 as well as words of caution.
1086 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1090 * All of the low level HMAC functions have been deprecated.
1092 *Paul Dale and David von Oheimb*
1094 * Over two thousand fixes were made to the documentation, including:
1095 - Common options (such as -rand/-writerand, TLS version control, etc)
1096 were refactored and point to newly-enhanced descriptions in openssl.pod.
1097 - Added style conformance for all options (with help from Richard Levitte),
1098 documented all reported missing options, added a CI build to check
1099 that all options are documented and that no unimplemented options
1101 - Documented some internals, such as all use of environment variables.
1102 - Addressed all internal broken L<> references.
1106 * All of the low level CMAC functions have been deprecated.
1110 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1111 functions have been deprecated.
1113 *Paul Dale and David von Oheimb*
1115 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1116 set of functions. The documentation mentioned negative values for some
1117 errors, but this was never the case, so the mention of negative values
1120 Code that followed the documentation and thereby check with something
1121 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1125 * All of the low level cipher functions have been deprecated.
1127 *Matt Caswell and Paul Dale*
1129 * Removed include/openssl/opensslconf.h.in and replaced it with
1130 include/openssl/configuration.h.in, which differs in not including
1131 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1132 was added to include both.
1134 This allows internal hacks where one might need to modify the set
1135 of configured macros, for example this if deprecated symbols are
1136 still supposed to be available internally:
1138 #include <openssl/configuration.h>
1140 #undef OPENSSL_NO_DEPRECATED
1141 #define OPENSSL_SUPPRESS_DEPRECATED
1143 #include <openssl/macros.h>
1145 This should not be used by applications that use the exported
1146 symbols, as that will lead to linking errors.
1150 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1151 used in exponentiation with 512-bit moduli. No EC algorithms are
1152 affected. Analysis suggests that attacks against 2-prime RSA1024,
1153 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1154 difficult to perform and are not believed likely. Attacks against DH512
1155 are considered just feasible. However, for an attack the target would
1156 have to re-use the DH512 private key, which is not recommended anyway.
1157 Also applications directly using the low-level API BN_mod_exp may be
1158 affected if they use BN_FLG_CONSTTIME.
1163 * Most memory-debug features have been deprecated, and the functionality
1164 replaced with no-ops.
1168 * Added documentation for the STACK API.
1172 * Introduced a new method type and API, OSSL_ENCODER, to represent
1173 generic encoders. These do the same sort of job that PEM writers
1174 and d2i functions do, but with support for methods supplied by
1175 providers, and the possibility for providers to support other
1180 * Introduced a new method type and API, OSSL_DECODER, to represent
1181 generic decoders. These do the same sort of job that PEM readers
1182 and i2d functions do, but with support for methods supplied by
1183 providers, and the possibility for providers to support other
1188 * Added a .pragma directive to the syntax of configuration files, to
1189 allow varying behavior in a supported and predictable manner.
1190 Currently added pragma:
1194 This allows dollar signs to be a keyword character unless it's
1195 followed by a opening brace or parenthesis. This is useful for
1196 platforms where dollar signs are commonly used in names, such as
1197 volume names and system directory names on VMS.
1201 * Added functionality to create an EVP_PKEY from user data.
1205 * Change the interpretation of the '--api' configuration option to
1206 mean that this is a desired API compatibility level with no
1207 further meaning. The previous interpretation, that this would
1208 also mean to remove all deprecated symbols up to and including
1209 the given version, no requires that 'no-deprecated' is also used
1210 in the configuration.
1212 When building applications, the desired API compatibility level
1213 can be set with the OPENSSL_API_COMPAT macro like before. For
1214 API compatibility version below 3.0, the old style numerical
1215 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1216 For version 3.0 and on, the value is expected to be the decimal
1217 value calculated from the major and minor version like this:
1219 MAJOR * 10000 + MINOR * 100
1223 -DOPENSSL_API_COMPAT=30000 For 3.0
1224 -DOPENSSL_API_COMPAT=30200 For 3.2
1226 To hide declarations that are deprecated up to and including the
1227 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1228 given when building the application as well.
1232 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1233 access to certificate and CRL stores via URIs and OSSL_STORE
1236 This adds the following functions:
1238 - X509_LOOKUP_store()
1239 - X509_STORE_load_file()
1240 - X509_STORE_load_path()
1241 - X509_STORE_load_store()
1242 - SSL_add_store_cert_subjects_to_stack()
1243 - SSL_CTX_set_default_verify_store()
1244 - SSL_CTX_load_verify_file()
1245 - SSL_CTX_load_verify_dir()
1246 - SSL_CTX_load_verify_store()
1250 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1251 The presence of this system service is determined at run-time.
1255 * Added functionality to create an EVP_PKEY context based on data
1256 for methods from providers. This takes an algorithm name and a
1257 property query string and simply stores them, with the intent
1258 that any operation that uses this context will use those strings
1259 to fetch the needed methods implicitly, thereby making the port
1260 of application written for pre-3.0 OpenSSL easier.
1264 * The undocumented function NCONF_WIN32() has been deprecated; for
1265 conversion details see the HISTORY section of doc/man5/config.pod
1269 * Introduced the new functions EVP_DigestSignInit_ex() and
1270 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1271 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1272 pages for further details.
1276 * Over two thousand fixes were made to the documentation, including:
1277 adding missing command flags, better style conformance, documentation
1280 *Rich Salz, Richard Levitte*
1282 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1283 X25519, X448, Ed25519 and Ed448.
1287 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1292 * Deprecated the public definition of `ERR_STATE` as well as the function
1293 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1298 * Added ERR functionality to give callers access to the stored function
1299 names that have replaced the older function code based functions.
1301 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1302 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1303 ERR_peek_error_all() and ERR_peek_last_error_all().
1305 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1306 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1307 ERR_func_error_string().
1311 * Extended testing to be verbose for failing tests only. The make variables
1312 VERBOSE_FAILURE or VF can be used to enable this:
1314 $ make VF=1 test # Unix
1315 $ mms /macro=(VF=1) test ! OpenVMS
1316 $ nmake VF=1 test # Windows
1320 * Added the `-copy_extensions` option to the `x509` command for use with
1321 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1322 all extensions in the request are copied to the certificate or vice versa.
1324 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1326 * Added the `-copy_extensions` option to the `req` command for use with
1327 `-x509`. When given with the `copy` or `copyall` argument,
1328 all extensions in the certification request are copied to the certificate.
1332 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1333 they generate are by default RFC 5280 compliant in the following sense:
1334 There is a subjectKeyIdentifier extension with a hash value of the public key
1335 and for not self-signed certs there is an authorityKeyIdentifier extension
1336 with a keyIdentifier field or issuer information identifying the signing key.
1337 This is done unless some configuration overrides the new default behavior,
1338 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1342 * Added several checks to `X509_verify_cert()` according to requirements in
1343 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1344 (which may be done by using the CLI option `-x509_strict`):
1345 * The basicConstraints of CA certificates must be marked critical.
1346 * CA certificates must explicitly include the keyUsage extension.
1347 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1348 * The issuer name of any certificate must not be empty.
1349 * The subject name of CA certs, certs with keyUsage crlSign,
1350 and certs without subjectAlternativeName must not be empty.
1351 * If a subjectAlternativeName extension is given it must not be empty.
1352 * The signatureAlgorithm field and the cert signature must be consistent.
1353 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1354 must not be marked critical.
1355 * The authorityKeyIdentifier must be given for X.509v3 certs
1356 unless they are self-signed.
1357 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1361 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1362 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1366 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1367 used even when parsing explicit parameters, when loading a encoded key
1368 or calling `EC_GROUP_new_from_ecpkparameters()`/
1369 `EC_GROUP_new_from_ecparameters()`.
1370 This prevents bypass of security hardening and performance gains,
1371 especially for curves with specialized EC_METHODs.
1372 By default, if a key encoded with explicit parameters is loaded and later
1373 encoded, the output is still encoded with explicit parameters, even if
1374 internally a "named" EC_GROUP is used for computation.
1378 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1379 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1380 NULL. After this change, only the cofactor parameter can be NULL. It also
1381 does some minimal sanity checks on the passed order.
1386 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1387 An attack is simple, if the first CMS_recipientInfo is valid but the
1388 second CMS_recipientInfo is chosen ciphertext. If the second
1389 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1390 encryption key will be replaced by garbage, and the message cannot be
1391 decoded, but if the RSA decryption fails, the correct encryption key is
1392 used and the recipient will not notice the attack.
1393 As a work around for this potential attack the length of the decrypted
1394 key must be equal to the cipher default key length, in case the
1395 certifiate is not given and all recipientInfo are tried out.
1396 The old behaviour can be re-enabled in the CMS code by setting the
1397 CMS_DEBUG_DECRYPT flag.
1401 * Early start up entropy quality from the DEVRANDOM seed source has been
1402 improved for older Linux systems. The RAND subsystem will wait for
1403 /dev/random to be producing output before seeding from /dev/urandom.
1404 The seeded state is stored for future library initialisations using
1405 a system global shared memory segment. The shared memory identifier
1406 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1407 the desired value. The default identifier is 114.
1411 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1412 when primes for RSA keys are computed.
1413 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1414 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1415 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1416 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1417 This avoids possible fingerprinting of newly generated RSA modules.
1421 * Correct the extended master secret constant on EBCDIC systems. Without this
1422 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1423 negotiate EMS will fail. Unfortunately this also means that TLS connections
1424 between EBCDIC systems with this fix, and EBCDIC systems without this
1425 fix will fail if they negotiate EMS.
1429 * Changed the library initialisation so that the config file is now loaded
1430 by default. This was already the case for libssl. It now occurs for both
1431 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1432 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1436 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1437 where the former acts as a replacement for `ERR_put_error()`, and the
1438 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1439 `ERR_raise_data()` adds more flexibility by taking a format string and
1440 an arbitrary number of arguments following it, to be processed with
1445 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1446 to check if a named provider is loaded and available. When called, it
1447 will also activate all fallback providers if such are still present.
1451 * Enforce a minimum DH modulus size of 512 bits.
1455 * Changed DH parameters to generate the order q subgroup instead of 2q.
1456 Previously generated DH parameters are still accepted by DH_check
1457 but DH_generate_key works around that by clearing bit 0 of the
1458 private key for those. This avoids leaking bit 0 of the private key.
1462 * Significantly reduce secure memory usage by the randomness pools.
1466 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1471 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1472 algorithms. An implementation of a key exchange algorithm can be obtained
1473 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1474 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1475 the older EVP_PKEY_derive_init() function. See the man pages for the new
1476 functions for further details.
1480 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1484 * Removed the function names from error messages and deprecated the
1489 * Removed NextStep support and the macro OPENSSL_UNISTD
1493 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1494 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1495 Also removed "export var as function" capability; we do not export
1496 variables, only functions.
1500 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1501 an error and 1 indicating success. In previous versions of OpenSSL this
1502 was a void type. If a key was set longer than the maximum possible this
1507 * Support SM2 signing and verification schemes with X509 certificate.
1511 * Use SHA256 as the default digest for TS query in the `ts` app.
1515 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1519 * Default cipher lists/suites are now available via a function, the
1520 #defines are deprecated.
1524 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1525 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1526 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1530 * Join the directories crypto/x509 and crypto/x509v3
1534 * Added command 'openssl kdf' that uses the EVP_KDF API.
1538 * Added command 'openssl mac' that uses the EVP_MAC API.
1542 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1543 as default directories. Also added the command 'openssl info'
1544 for scripting purposes.
1548 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1553 * Add prediction resistance to the DRBG reseeding process.
1557 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1558 mandated by IEEE Std 1619-2018.
1562 * Added newline escaping functionality to a filename when using openssl dgst.
1563 This output format is to replicate the output format found in the `*sum`
1564 checksum programs. This aims to preserve backward compatibility.
1566 *Matt Eaton, Richard Levitte, and Paul Dale*
1568 * Removed the heartbeat message in DTLS feature, as it has very
1569 little usage and doesn't seem to fulfill a valuable purpose.
1570 The configuration option is now deprecated.
1574 * Changed the output of 'openssl {digestname} < file' to display the
1575 digest name in its output.
1579 * Added a new generic trace API which provides support for enabling
1580 instrumentation through trace output.
1582 *Richard Levitte & Matthias St. Pierre*
1584 * Added build tests for C++. These are generated files that only do one
1585 thing, to include one public OpenSSL head file each. This tests that
1586 the public header files can be usefully included in a C++ application.
1588 This test isn't enabled by default. It can be enabled with the option
1589 'enable-buildtest-c++'.
1593 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1597 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1601 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1605 * Added KMAC to EVP_MAC.
1609 * Added property based algorithm implementation selection framework to
1614 * Added SCA hardening for modular field inversion in EC_GROUP through
1615 a new dedicated field_inv() pointer in EC_METHOD.
1616 This also addresses a leakage affecting conversions from projective
1617 to affine coordinates.
1619 *Billy Bob Brumley, Nicola Tuveri*
1621 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1622 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1623 those algorithms that were already supported through the EVP_PKEY API
1624 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1625 and scrypt are now wrappers that call EVP_KDF.
1629 * Build devcrypto engine as a dynamic engine.
1631 *Eneas U de Queiroz*
1633 * Add keyed BLAKE2 to EVP_MAC.
1637 * Fix a bug in the computation of the endpoint-pair shared secret used
1638 by DTLS over SCTP. This breaks interoperability with older versions
1639 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1640 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1641 interoperability with such broken implementations. However, enabling
1642 this switch breaks interoperability with correct implementations.
1644 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1645 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1649 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1653 * Changed the license to the Apache License v2.0.
1657 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1659 - Major releases (indicated by incrementing the MAJOR release number)
1660 may introduce incompatible API/ABI changes.
1661 - Minor releases (indicated by incrementing the MINOR release number)
1662 may introduce new features but retain API/ABI compatibility.
1663 - Patch releases (indicated by incrementing the PATCH number)
1664 are intended for bug fixes and other improvements of existing
1665 features only (like improving performance or adding documentation)
1666 and retain API/ABI compatibility.
1670 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1674 * Remove the 'dist' target and add a tarball building script. The
1675 'dist' target has fallen out of use, and it shouldn't be
1676 necessary to configure just to create a source distribution.
1680 * Recreate the OS390-Unix config target. It no longer relies on a
1681 special script like it did for OpenSSL pre-1.1.0.
1685 * Instead of having the source directories listed in Configure, add
1686 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1691 * Add GMAC to EVP_MAC.
1695 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1699 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1700 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1701 to facilitate the continued use of MACs through raw private keys in
1702 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1706 * Deprecate ECDH_KDF_X9_62().
1710 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1711 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1712 are retained for backwards compatibility.
1716 * AES-XTS mode now enforces that its two keys are different to mitigate
1717 the attacked described in "Efficient Instantiations of Tweakable
1718 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1719 Details of this attack can be obtained from:
1720 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1724 * Rename the object files, i.e. give them other names than in previous
1725 versions. Their names now include the name of the final product, as
1726 well as its type mnemonic (bin, lib, shlib).
1730 * Added new option for 'openssl list', '-objects', which will display the
1731 list of built in objects, i.e. OIDs with names.
1735 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1736 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1741 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1742 improves application performance by removing data copies and providing
1743 applications with zero-copy system calls such as sendfile and splice.
1747 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1751 * `PKCS12_parse` now maintains the order of the parsed certificates
1752 when outputting them via `*ca` (rather than reversing it).
1756 * Deprecated pthread fork support methods.
1760 * Added support for FFDHE key exchange in TLS 1.3.
1764 * Added a new concept for OpenSSL plugability: providers. This
1765 functionality is designed to replace the ENGINE API and ENGINE
1766 implementations, and to be much more dynamic, allowing provider
1767 authors to introduce new algorithms among other things, as long as
1768 there's an API that supports the algorithm type.
1770 With this concept comes a new core API for interaction between
1771 libcrypto and provider implementations. Public libcrypto functions
1772 that want to use providers do so through this core API.
1774 The main documentation for this core API is found in
1775 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1776 refer to other manuals describing the API specific for supported
1777 algorithm types (also called operations).
1784 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1786 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
1788 * Avoid loading of a dynamic engine twice.
1792 * Prioritise DANE TLSA issuer certs over peer certs
1796 * Fixed random API for MacOS prior to 10.12
1798 These MacOS versions don't support the CommonCrypto APIs
1802 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1804 * Fixed an SM2 Decryption Buffer Overflow.
1806 In order to decrypt SM2 encrypted data an application is expected to
1807 call the API function EVP_PKEY_decrypt(). Typically an application will
1808 call this function twice. The first time, on entry, the "out" parameter
1809 can be NULL and, on exit, the "outlen" parameter is populated with the
1810 buffer size required to hold the decrypted plaintext. The application
1811 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1812 again, but this time passing a non-NULL value for the "out" parameter.
1814 A bug in the implementation of the SM2 decryption code means that the
1815 calculation of the buffer size required to hold the plaintext returned
1816 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1817 size required by the second call. This can lead to a buffer overflow
1818 when EVP_PKEY_decrypt() is called by the application a second time with
1819 a buffer that is too small.
1821 A malicious attacker who is able present SM2 content for decryption to
1822 an application could cause attacker chosen data to overflow the buffer
1823 by up to a maximum of 62 bytes altering the contents of other data held
1824 after the buffer, possibly changing application behaviour or causing
1825 the application to crash. The location of the buffer is application
1826 dependent but is typically heap allocated.
1831 * Fixed various read buffer overruns processing ASN.1 strings
1833 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1834 structure which contains a buffer holding the string data and a field
1835 holding the buffer length. This contrasts with normal C strings which
1836 are repesented as a buffer for the string data which is terminated
1837 with a NUL (0) byte.
1839 Although not a strict requirement, ASN.1 strings that are parsed using
1840 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1841 well as any string whose value has been set with the ASN1_STRING_set()
1842 function will additionally NUL terminate the byte array in the
1843 ASN1_STRING structure.
1845 However, it is possible for applications to directly construct valid
1846 ASN1_STRING structures which do not NUL terminate the byte array by
1847 directly setting the "data" and "length" fields in the ASN1_STRING
1848 array. This can also happen by using the ASN1_STRING_set0() function.
1850 Numerous OpenSSL functions that print ASN.1 data have been found to
1851 assume that the ASN1_STRING byte array will be NUL terminated, even
1852 though this is not guaranteed for strings that have been directly
1853 constructed. Where an application requests an ASN.1 structure to be
1854 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1855 been directly constructed by the application without NUL terminating
1856 the "data" field, then a read buffer overrun can occur.
1858 The same thing can also occur during name constraints processing
1859 of certificates (for example if a certificate has been directly
1860 constructed by the application instead of loading it via the OpenSSL
1861 parsing functions, and the certificate contains non NUL terminated
1862 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1863 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1865 If a malicious actor can cause an application to directly construct an
1866 ASN1_STRING and then process it through one of the affected OpenSSL
1867 functions then this issue could be hit. This might result in a crash
1868 (causing a Denial of Service attack). It could also result in the
1869 disclosure of private memory contents (such as private keys, or
1870 sensitive plaintext).
1875 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1877 * Fixed a problem with verifying a certificate chain when using the
1878 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1879 the certificates present in a certificate chain. It is not set by default.
1881 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1882 the chain that have explicitly encoded elliptic curve parameters was added
1883 as an additional strict check.
1885 An error in the implementation of this check meant that the result of a
1886 previous check to confirm that certificates in the chain are valid CA
1887 certificates was overwritten. This effectively bypasses the check
1888 that non-CA certificates must not be able to issue other certificates.
1890 If a "purpose" has been configured then there is a subsequent opportunity
1891 for checks that the certificate is a valid CA. All of the named "purpose"
1892 values implemented in libcrypto perform this check. Therefore, where
1893 a purpose is set the certificate chain will still be rejected even when the
1894 strict flag has been used. A purpose is set by default in libssl client and
1895 server certificate verification routines, but it can be overridden or
1896 removed by an application.
1898 In order to be affected, an application must explicitly set the
1899 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1900 for the certificate verification or, in the case of TLS client or server
1901 applications, override the default purpose.
1906 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1907 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1908 renegotiation ClientHello omits the signature_algorithms extension (where it
1909 was present in the initial ClientHello), but includes a
1910 signature_algorithms_cert extension then a NULL pointer dereference will
1911 result, leading to a crash and a denial of service attack.
1913 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1914 (which is the default configuration). OpenSSL TLS clients are not impacted by
1918 *Peter Kästle and Samuel Sapalski*
1920 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1922 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1923 create a unique hash value based on the issuer and serial number data
1924 contained within an X509 certificate. However it was failing to correctly
1925 handle any errors that may occur while parsing the issuer field (which might
1926 occur if the issuer field is maliciously constructed). This may subsequently
1927 result in a NULL pointer deref and a crash leading to a potential denial of
1933 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1934 padding mode to correctly check for rollback attacks. This is considered a
1935 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1940 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1941 functions. Previously they could overflow the output length argument in some
1942 cases where the input length is close to the maximum permissable length for
1943 an integer on the platform. In such cases the return value from the function
1944 call would be 1 (indicating success), but the output length value would be
1945 negative. This could cause applications to behave incorrectly or crash.
1950 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1951 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1952 could be exploited in a side channel attack to recover the password. Since
1953 the attack is local host only this is outside of the current OpenSSL
1954 threat model and therefore no CVE is assigned.
1956 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1961 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1963 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1964 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1965 If an attacker can control both items being compared then this could lead
1966 to a possible denial of service attack. OpenSSL itself uses the
1967 GENERAL_NAME_cmp function for two purposes:
1968 1) Comparing CRL distribution point names between an available CRL and a
1969 CRL distribution point embedded in an X509 certificate
1970 2) When verifying that a timestamp response token signer matches the
1971 timestamp authority name (exposed via the API functions
1972 TS_RESP_verify_response and TS_RESP_verify_token)
1977 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1979 * Certificates with explicit curve parameters are now disallowed in
1980 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1984 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1985 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1986 conversely, silently ignore DTLS protocol version bounds when configuring
1987 TLS-based contexts. The commands can be repeated to set bounds of both
1988 types. The same applies with the corresponding "min_protocol" and
1989 "max_protocol" command-line switches, in case some application uses both TLS
1992 SSL_CTX instances that are created for a fixed protocol version (e.g.
1993 TLSv1_server_method()) also silently ignore version bounds. Previously
1994 attempts to apply bounds to these protocol versions would result in an
1995 error. Now only the "version-flexible" SSL_CTX instances are subject to
1996 limits in configuration files in command-line options.
2000 * Handshake now fails if Extended Master Secret extension is dropped
2005 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2007 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2009 * Fixed segmentation fault in SSL_check_chain()
2010 Server or client applications that call the SSL_check_chain() function
2011 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2012 dereference as a result of incorrect handling of the
2013 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2014 or unrecognised signature algorithm is received from the peer. This could
2015 be exploited by a malicious peer in a Denial of Service attack.
2020 * Added AES consttime code for no-asm configurations
2021 an optional constant time support for AES was added
2022 when building openssl for no-asm.
2023 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2024 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2025 At this time this feature is by default disabled.
2026 It will be enabled by default in 3.0.
2030 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2032 * Revert the change of EOF detection while reading in libssl to avoid
2033 regressions in applications depending on the current way of reporting
2034 the EOF. As the existing method is not fully accurate the change to
2035 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2036 branch and will be present in the 3.0 release.
2040 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2041 when primes for RSA keys are computed.
2042 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2043 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2044 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2045 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2046 This avoids possible fingerprinting of newly generated RSA modules.
2050 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2052 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2053 while reading in libssl then we would report an error back to the
2054 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2055 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2056 therefore give a hint as to what went wrong.
2060 * Check that ed25519 and ed448 are allowed by the security level. Previously
2061 signature algorithms not using an MD were not being checked that they were
2062 allowed by the security level.
2066 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2067 was not quite right. The behaviour was not consistent between resumption
2068 and normal handshakes, and also not quite consistent with historical
2069 behaviour. The behaviour in various scenarios has been clarified and
2070 it has been updated to make it match historical behaviour as closely as
2075 * *[VMS only]* The header files that the VMS compilers include automatically,
2076 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2077 that the C++ compiler doesn't understand. This is a shortcoming in the
2078 compiler, but can be worked around with `__cplusplus` guards.
2080 C++ applications that use OpenSSL libraries must be compiled using the
2081 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2082 functions. Otherwise, only functions with symbols of less than 31
2083 characters can be used, as the linker will not be able to successfully
2084 resolve symbols with longer names.
2088 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2089 The presence of this system service is determined at run-time.
2093 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2098 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2100 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2101 number generator (RNG). This was intended to include protection in the
2102 event of a fork() system call in order to ensure that the parent and child
2103 processes did not share the same RNG state. However this protection was not
2104 being used in the default case.
2106 A partial mitigation for this issue is that the output from a high
2107 precision timer is mixed into the RNG state so the likelihood of a parent
2108 and child process sharing state is significantly reduced.
2110 If an application already calls OPENSSL_init_crypto() explicitly using
2111 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2114 *Matthias St. Pierre*
2116 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2117 used even when parsing explicit parameters, when loading a encoded key
2118 or calling `EC_GROUP_new_from_ecpkparameters()`/
2119 `EC_GROUP_new_from_ecparameters()`.
2120 This prevents bypass of security hardening and performance gains,
2121 especially for curves with specialized EC_METHODs.
2122 By default, if a key encoded with explicit parameters is loaded and later
2123 encoded, the output is still encoded with explicit parameters, even if
2124 internally a "named" EC_GROUP is used for computation.
2128 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2129 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2130 NULL. After this change, only the cofactor parameter can be NULL. It also
2131 does some minimal sanity checks on the passed order.
2136 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2137 An attack is simple, if the first CMS_recipientInfo is valid but the
2138 second CMS_recipientInfo is chosen ciphertext. If the second
2139 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2140 encryption key will be replaced by garbage, and the message cannot be
2141 decoded, but if the RSA decryption fails, the correct encryption key is
2142 used and the recipient will not notice the attack.
2143 As a work around for this potential attack the length of the decrypted
2144 key must be equal to the cipher default key length, in case the
2145 certifiate is not given and all recipientInfo are tried out.
2146 The old behaviour can be re-enabled in the CMS code by setting the
2147 CMS_DEBUG_DECRYPT flag.
2152 * Early start up entropy quality from the DEVRANDOM seed source has been
2153 improved for older Linux systems. The RAND subsystem will wait for
2154 /dev/random to be producing output before seeding from /dev/urandom.
2155 The seeded state is stored for future library initialisations using
2156 a system global shared memory segment. The shared memory identifier
2157 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2158 the desired value. The default identifier is 114.
2162 * Correct the extended master secret constant on EBCDIC systems. Without this
2163 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2164 negotiate EMS will fail. Unfortunately this also means that TLS connections
2165 between EBCDIC systems with this fix, and EBCDIC systems without this
2166 fix will fail if they negotiate EMS.
2170 * Use Windows installation paths in the mingw builds
2172 Mingw isn't a POSIX environment per se, which means that Windows
2173 paths should be used for installation.
2178 * Changed DH_check to accept parameters with order q and 2q subgroups.
2179 With order 2q subgroups the bit 0 of the private key is not secret
2180 but DH_generate_key works around that by clearing bit 0 of the
2181 private key for those. This avoids leaking bit 0 of the private key.
2185 * Significantly reduce secure memory usage by the randomness pools.
2189 * Revert the DEVRANDOM_WAIT feature for Linux systems
2191 The DEVRANDOM_WAIT feature added a select() call to wait for the
2192 /dev/random device to become readable before reading from the
2193 /dev/urandom device.
2195 It turned out that this change had negative side effects on
2196 performance which were not acceptable. After some discussion it
2197 was decided to revert this feature and leave it up to the OS
2198 resp. the platform maintainer to ensure a proper initialization
2199 during early boot time.
2201 *Matthias St. Pierre*
2203 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2205 * Add build tests for C++. These are generated files that only do one
2206 thing, to include one public OpenSSL head file each. This tests that
2207 the public header files can be usefully included in a C++ application.
2209 This test isn't enabled by default. It can be enabled with the option
2210 'enable-buildtest-c++'.
2214 * Enable SHA3 pre-hashing for ECDSA and DSA.
2218 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2219 This changes the size when using the `genpkey` command when no size is given.
2220 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2221 generation commands to use 2048 bits by default.
2225 * Reorganize the manual pages to consistently have RETURN VALUES,
2226 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2227 util/fix-doc-nits accordingly.
2229 *Paul Yang, Joshua Lock*
2231 * Add the missing accessor EVP_PKEY_get0_engine()
2235 * Have commands like `s_client` and `s_server` output the signature scheme
2236 along with other cipher suite parameters when debugging.
2240 * Make OPENSSL_config() error agnostic again.
2244 * Do the error handling in RSA decryption constant time.
2248 * Prevent over long nonces in ChaCha20-Poly1305.
2250 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2251 for every encryption operation. RFC 7539 specifies that the nonce value
2252 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2253 and front pads the nonce with 0 bytes if it is less than 12
2254 bytes. However it also incorrectly allows a nonce to be set of up to 16
2255 bytes. In this case only the last 12 bytes are significant and any
2256 additional leading bytes are ignored.
2258 It is a requirement of using this cipher that nonce values are
2259 unique. Messages encrypted using a reused nonce value are susceptible to
2260 serious confidentiality and integrity attacks. If an application changes
2261 the default nonce length to be longer than 12 bytes and then makes a
2262 change to the leading bytes of the nonce expecting the new value to be a
2263 new unique nonce then such an application could inadvertently encrypt
2264 messages with a reused nonce.
2266 Additionally the ignored bytes in a long nonce are not covered by the
2267 integrity guarantee of this cipher. Any application that relies on the
2268 integrity of these ignored leading bytes of a long nonce may be further
2269 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2270 is safe because no such use sets such a long nonce value. However user
2271 applications that use this cipher directly and set a non-default nonce
2272 length to be longer than 12 bytes may be vulnerable.
2274 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2280 * Add DEVRANDOM_WAIT feature for Linux systems
2282 On older Linux systems where the getrandom() system call is not available,
2283 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2284 Contrary to getrandom(), the /dev/urandom device will not block during
2285 early boot when the kernel CSPRNG has not been seeded yet.
2287 To mitigate this known weakness, use select() to wait for /dev/random to
2288 become readable before reading from /dev/urandom.
2290 * Ensure that SM2 only uses SM3 as digest algorithm
2294 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2296 * Change the info callback signals for the start and end of a post-handshake
2297 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2298 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2299 confused by this and assume that a TLSv1.2 renegotiation has started. This
2300 can break KeyUpdate handling. Instead we no longer signal the start and end
2301 of a post handshake message exchange (although the messages themselves are
2302 still signalled). This could break some applications that were expecting
2303 the old signals. However without this KeyUpdate is not usable for many
2308 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2310 * Timing vulnerability in DSA signature generation
2312 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2313 timing side channel attack. An attacker could use variations in the signing
2314 algorithm to recover the private key.
2316 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2321 * Timing vulnerability in ECDSA signature generation
2323 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2324 timing side channel attack. An attacker could use variations in the signing
2325 algorithm to recover the private key.
2327 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2332 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2333 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2334 of two gigabytes and the error handling improved.
2336 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2337 categorized as a normal bug, not a security issue, because the DRBG reseeds
2338 automatically and is fully functional even without additional randomness
2339 provided by the application.
2341 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2343 * Add a new ClientHello callback. Provides a callback interface that gives
2344 the application the ability to adjust the nascent SSL object at the
2345 earliest stage of ClientHello processing, immediately after extensions have
2346 been collected but before they have been processed. In particular, this
2347 callback can adjust the supported TLS versions in response to the contents
2352 * Add SM2 base algorithm support.
2356 * s390x assembly pack: add (improved) hardware-support for the following
2357 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2358 aes-cfb/cfb8, aes-ecb.
2362 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2363 parameter is no longer accepted, as it leads to a corrupt table. NULL
2364 pem_str is reserved for alias entries only.
2368 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2369 step for prime curves. The new implementation is based on formulae from
2370 differential addition-and-doubling in homogeneous projective coordinates
2371 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2372 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2373 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2374 to work in projective coordinates.
2376 *Billy Bob Brumley, Nicola Tuveri*
2378 * Change generating and checking of primes so that the error rate of not
2379 being prime depends on the intended use based on the size of the input.
2380 For larger primes this will result in more rounds of Miller-Rabin.
2381 The maximal error rate for primes with more than 1080 bits is lowered
2384 *Kurt Roeckx, Annie Yousar*
2386 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2390 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2391 moving between systems, and to avoid confusion when a Windows build is
2392 done with mingw vs with MSVC. For POSIX installs, there's still a
2393 symlink or copy named 'tsget' to avoid that confusion as well.
2397 * Revert blinding in ECDSA sign and instead make problematic addition
2398 length-invariant. Switch even to fixed-length Montgomery multiplication.
2402 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2403 step for binary curves. The new implementation is based on formulae from
2404 differential addition-and-doubling in mixed Lopez-Dahab projective
2405 coordinates, modified to independently blind the operands.
2407 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2409 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2410 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2411 EC_METHODs to implement their own specialized "ladder step", to take
2412 advantage of more favorable coordinate systems or more efficient
2413 differential addition-and-doubling algorithms.
2415 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2417 * Modified the random device based seed sources to keep the relevant
2418 file descriptors open rather than reopening them on each access.
2419 This allows such sources to operate in a chroot() jail without
2420 the associated device nodes being available. This behaviour can be
2421 controlled using RAND_keep_random_devices_open().
2425 * Numerous side-channel attack mitigations have been applied. This may have
2426 performance impacts for some algorithms for the benefit of improved
2427 security. Specific changes are noted in this change log by their respective
2432 * AIX shared library support overhaul. Switch to AIX "natural" way of
2433 handling shared libraries, which means collecting shared objects of
2434 different versions and bitnesses in one common archive. This allows to
2435 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2436 doesn't affect the way 3rd party applications are linked, only how
2437 multi-version installation is managed.
2441 * Make ec_group_do_inverse_ord() more robust and available to other
2442 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2443 mitigations are applied to the fallback BN_mod_inverse().
2444 When using this function rather than BN_mod_inverse() directly, new
2445 EC cryptosystem implementations are then safer-by-default.
2449 * Add coordinate blinding for EC_POINT and implement projective
2450 coordinate blinding for generic prime curves as a countermeasure to
2451 chosen point SCA attacks.
2453 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2455 * Add blinding to ECDSA and DSA signatures to protect against side channel
2456 attacks discovered by Keegan Ryan (NCC Group).
2460 * Enforce checking in the `pkeyutl` command to ensure that the input
2461 length does not exceed the maximum supported digest length when performing
2462 a sign, verify or verifyrecover operation.
2466 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2467 I/O in combination with something like select() or poll() will hang. This
2468 can be turned off again using SSL_CTX_clear_mode().
2469 Many applications do not properly handle non-application data records, and
2470 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2471 around the problems in those applications, but can also break some.
2472 It's recommended to read the manpages about SSL_read(), SSL_write(),
2473 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2474 SSL_CTX_set_read_ahead() again.
2478 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2479 now allow empty (zero character) pass phrases.
2483 * Apply blinding to binary field modular inversion and remove patent
2484 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2488 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2489 binary and prime elliptic curves.
2493 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2494 constant time fixed point multiplication.
2498 * Revise elliptic curve scalar multiplication with timing attack
2499 defenses: ec_wNAF_mul redirects to a constant time implementation
2500 when computing fixed point and variable point multiplication (which
2501 in OpenSSL are mostly used with secret scalars in keygen, sign,
2502 ECDH derive operations).
2503 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2506 * Updated CONTRIBUTING
2510 * Updated DRBG / RAND to request nonce and additional low entropy
2511 randomness from the system.
2513 *Matthias St. Pierre*
2515 * Updated 'openssl rehash' to use OpenSSL consistent default.
2519 * Moved the load of the ssl_conf module to libcrypto, which helps
2520 loading engines that libssl uses before libssl is initialised.
2524 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2528 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2530 *Ingo Schwarze, Rich Salz*
2532 * Added output of accepting IP address and port for 'openssl s_server'
2536 * Added a new API for TLSv1.3 ciphersuites:
2537 SSL_CTX_set_ciphersuites()
2538 SSL_set_ciphersuites()
2542 * Memory allocation failures consistently add an error to the error
2547 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2548 in libcrypto when run as setuid/setgid.
2552 * Load any config file by default when libssl is used.
2556 * Added new public header file <openssl/rand_drbg.h> and documentation
2557 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2559 *Matthias St. Pierre*
2561 * QNX support removed (cannot find contributors to get their approval
2562 for the license change).
2566 * TLSv1.3 replay protection for early data has been implemented. See the
2567 SSL_read_early_data() man page for further details.
2571 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2572 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2573 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2574 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2575 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2576 configuration has been separated out. See the ciphers man page or the
2577 SSL_CTX_set_ciphersuites() man page for more information.
2581 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2582 in responder mode now supports the new "-multi" option, which
2583 spawns the specified number of child processes to handle OCSP
2584 requests. The "-timeout" option now also limits the OCSP
2585 responder's patience to wait to receive the full client request
2586 on a newly accepted connection. Child processes are respawned
2587 as needed, and the CA index file is automatically reloaded
2588 when changed. This makes it possible to run the "ocsp" responder
2589 as a long-running service, making the OpenSSL CA somewhat more
2590 feature-complete. In this mode, most diagnostic messages logged
2591 after entering the event loop are logged via syslog(3) rather than
2596 * Added support for X448 and Ed448. Heavily based on original work by
2601 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2602 objects loaded. This adds the functions OSSL_STORE_expect() and
2603 OSSL_STORE_find() as well as needed tools to construct searches and
2604 get the search data out of them.
2608 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2609 version of OpenSSL should review their configuration settings to ensure
2610 that they are still appropriate for TLSv1.3. For further information see:
2611 <https://wiki.openssl.org/index.php/TLS1.3>
2615 * Grand redesign of the OpenSSL random generator
2617 The default RAND method now utilizes an AES-CTR DRBG according to
2618 NIST standard SP 800-90Ar1. The new random generator is essentially
2619 a port of the default random generator from the OpenSSL FIPS 2.0
2620 object module. It is a hybrid deterministic random bit generator
2621 using an AES-CTR bit stream and which seeds and reseeds itself
2622 automatically using trusted system entropy sources.
2624 Some of its new features are:
2625 - Support for multiple DRBG instances with seed chaining.
2626 - The default RAND method makes use of a DRBG.
2627 - There is a public and private DRBG instance.
2628 - The DRBG instances are fork-safe.
2629 - Keep all global DRBG instances on the secure heap if it is enabled.
2630 - The public and private DRBG instance are per thread for lock free
2633 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2635 * Changed Configure so it only says what it does and doesn't dump
2636 so much data. Instead, ./configdata.pm should be used as a script
2637 to display all sorts of configuration data.
2641 * Added processing of "make variables" to Configure.
2645 * Added SHA512/224 and SHA512/256 algorithm support.
2649 * The last traces of Netware support, first removed in 1.1.0, have
2654 * Get rid of Makefile.shared, and in the process, make the processing
2655 of certain files (rc.obj, or the .def/.map/.opt files produced from
2656 the ordinal files) more visible and hopefully easier to trace and
2657 debug (or make silent).
2661 * Make it possible to have environment variable assignments as
2662 arguments to config / Configure.
2666 * Add multi-prime RSA (RFC 8017) support.
2670 * Add SM3 implemented according to GB/T 32905-2016
2671 *Jack Lloyd <jack.lloyd@ribose.com>,*
2672 *Ronald Tse <ronald.tse@ribose.com>,*
2673 *Erick Borsboom <erick.borsboom@ribose.com>*
2675 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2676 as documented in RFC6066.
2677 Based on a patch from Tomasz Moń
2679 *Filipe Raimundo da Silva*
2681 * Add SM4 implemented according to GB/T 32907-2016.
2682 *Jack Lloyd <jack.lloyd@ribose.com>,*
2683 *Ronald Tse <ronald.tse@ribose.com>,*
2684 *Erick Borsboom <erick.borsboom@ribose.com>*
2686 * Reimplement -newreq-nodes and ERR_error_string_n; the
2687 original author does not agree with the license change.
2691 * Add ARIA AEAD TLS support.
2695 * Some macro definitions to support VS6 have been removed. Visual
2696 Studio 6 has not worked since 1.1.0
2700 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2701 without clearing the errors.
2705 * Add "atfork" functions. If building on a system that without
2706 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2707 requirements. The RAND facility now uses/requires this.
2715 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2716 not possible to disable entirely. However, it's still possible to
2717 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2720 To disable, configure with 'no-ui-console'. 'no-ui' is still
2721 possible to use as an alias. Check at compile time with the
2722 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2723 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2727 * Add a STORE module, which implements a uniform and URI based reader of
2728 stores that can contain keys, certificates, CRLs and numerous other
2729 objects. The main API is loosely based on a few stdio functions,
2730 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2731 OSSL_STORE_error and OSSL_STORE_close.
2732 The implementation uses backends called "loaders" to implement arbitrary
2733 URI schemes. There is one built in "loader" for the 'file' scheme.
2737 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2738 then adjusted to work on FreeBSD 8.4 as well.
2739 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2740 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2744 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2745 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2746 error code calls like this:
2748 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2750 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2751 that can be encoded in C. For the foreseeable future, this will only
2754 *Richard Levitte and Tim Hudson*
2756 * Removed BSD cryptodev engine.
2760 * Add a build target 'build_all_generated', to build all generated files
2761 and only that. This can be used to prepare everything that requires
2762 things like perl for a system that lacks perl and then move everything
2763 to that system and do the rest of the build there.
2767 * In the UI interface, make it possible to duplicate the user data. This
2768 can be used by engines that need to retain the data for a longer time
2769 than just the call where this user data is passed.
2773 * Ignore the '-named_curve auto' value for compatibility of applications
2776 *Tomáš Mráz <tmraz@fedoraproject.org>*
2778 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2779 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2780 alerts across multiple records (some of which could be empty). In practice
2781 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2782 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2783 support this at all. Supporting it adds significant complexity to the
2784 record layer, and its removal is unlikely to cause interoperability
2789 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2790 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2791 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2796 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2797 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2799 *Richard Levitte, Andy Polyakov*
2801 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2806 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2807 platform rather than 'mingw'.
2811 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2812 success if they are asked to add an object which already exists
2813 in the store. This change cascades to other functions which load
2814 certificates and CRLs.
2818 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2819 facilitate stack unwinding even from assembly subroutines.
2823 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2824 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2828 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2829 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2830 which is the minimum version we support.
2834 * Certificate time validation (X509_cmp_time) enforces stricter
2835 compliance with RFC 5280. Fractional seconds and timezone offsets
2836 are no longer allowed.
2840 * Add support for ARIA
2844 * s_client will now send the Server Name Indication (SNI) extension by
2845 default unless the new "-noservername" option is used. The server name is
2846 based on the host provided to the "-connect" option unless overridden by
2847 using "-servername".
2851 * Add support for SipHash
2855 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2856 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2857 prevent issues where no progress is being made and the peer continually
2858 sends unrecognised record types, using up resources processing them.
2862 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2863 using the algorithm defined in
2864 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2868 * Heartbeat support has been removed; the ABI is changed for now.
2870 *Richard Levitte, Rich Salz*
2872 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2876 * The RSA "null" method, which was partially supported to avoid patent
2877 issues, has been replaced to always returns NULL.
2884 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2886 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2887 used even when parsing explicit parameters, when loading a encoded key
2888 or calling `EC_GROUP_new_from_ecpkparameters()`/
2889 `EC_GROUP_new_from_ecparameters()`.
2890 This prevents bypass of security hardening and performance gains,
2891 especially for curves with specialized EC_METHODs.
2892 By default, if a key encoded with explicit parameters is loaded and later
2893 encoded, the output is still encoded with explicit parameters, even if
2894 internally a "named" EC_GROUP is used for computation.
2898 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2899 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2900 NULL. After this change, only the cofactor parameter can be NULL. It also
2901 does some minimal sanity checks on the passed order.
2906 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2907 An attack is simple, if the first CMS_recipientInfo is valid but the
2908 second CMS_recipientInfo is chosen ciphertext. If the second
2909 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2910 encryption key will be replaced by garbage, and the message cannot be
2911 decoded, but if the RSA decryption fails, the correct encryption key is
2912 used and the recipient will not notice the attack.
2913 As a work around for this potential attack the length of the decrypted
2914 key must be equal to the cipher default key length, in case the
2915 certifiate is not given and all recipientInfo are tried out.
2916 The old behaviour can be re-enabled in the CMS code by setting the
2917 CMS_DEBUG_DECRYPT flag.
2922 * Use Windows installation paths in the mingw builds
2924 Mingw isn't a POSIX environment per se, which means that Windows
2925 paths should be used for installation.
2930 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2932 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2933 This changes the size when using the `genpkey` command when no size is given.
2934 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2935 generation commands to use 2048 bits by default.
2939 * Prevent over long nonces in ChaCha20-Poly1305.
2941 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2942 for every encryption operation. RFC 7539 specifies that the nonce value
2943 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2944 and front pads the nonce with 0 bytes if it is less than 12
2945 bytes. However it also incorrectly allows a nonce to be set of up to 16
2946 bytes. In this case only the last 12 bytes are significant and any
2947 additional leading bytes are ignored.
2949 It is a requirement of using this cipher that nonce values are
2950 unique. Messages encrypted using a reused nonce value are susceptible to
2951 serious confidentiality and integrity attacks. If an application changes
2952 the default nonce length to be longer than 12 bytes and then makes a
2953 change to the leading bytes of the nonce expecting the new value to be a
2954 new unique nonce then such an application could inadvertently encrypt
2955 messages with a reused nonce.
2957 Additionally the ignored bytes in a long nonce are not covered by the
2958 integrity guarantee of this cipher. Any application that relies on the
2959 integrity of these ignored leading bytes of a long nonce may be further
2960 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2961 is safe because no such use sets such a long nonce value. However user
2962 applications that use this cipher directly and set a non-default nonce
2963 length to be longer than 12 bytes may be vulnerable.
2965 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2971 * Added SCA hardening for modular field inversion in EC_GROUP through
2972 a new dedicated field_inv() pointer in EC_METHOD.
2973 This also addresses a leakage affecting conversions from projective
2974 to affine coordinates.
2976 *Billy Bob Brumley, Nicola Tuveri*
2978 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2979 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2983 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2987 * Remove the 'dist' target and add a tarball building script. The
2988 'dist' target has fallen out of use, and it shouldn't be
2989 necessary to configure just to create a source distribution.
2993 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2995 * Timing vulnerability in DSA signature generation
2997 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2998 timing side channel attack. An attacker could use variations in the signing
2999 algorithm to recover the private key.
3001 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3006 * Timing vulnerability in ECDSA signature generation
3008 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3009 timing side channel attack. An attacker could use variations in the signing
3010 algorithm to recover the private key.
3012 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3017 * Add coordinate blinding for EC_POINT and implement projective
3018 coordinate blinding for generic prime curves as a countermeasure to
3019 chosen point SCA attacks.
3021 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3023 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3025 * Client DoS due to large DH parameter
3027 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3028 malicious server can send a very large prime value to the client. This will
3029 cause the client to spend an unreasonably long period of time generating a
3030 key for this prime resulting in a hang until the client has finished. This
3031 could be exploited in a Denial Of Service attack.
3033 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3038 * Cache timing vulnerability in RSA Key Generation
3040 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3041 a cache timing side channel attack. An attacker with sufficient access to
3042 mount cache timing attacks during the RSA key generation process could
3043 recover the private key.
3045 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3046 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3051 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3052 parameter is no longer accepted, as it leads to a corrupt table. NULL
3053 pem_str is reserved for alias entries only.
3057 * Revert blinding in ECDSA sign and instead make problematic addition
3058 length-invariant. Switch even to fixed-length Montgomery multiplication.
3062 * Change generating and checking of primes so that the error rate of not
3063 being prime depends on the intended use based on the size of the input.
3064 For larger primes this will result in more rounds of Miller-Rabin.
3065 The maximal error rate for primes with more than 1080 bits is lowered
3068 *Kurt Roeckx, Annie Yousar*
3070 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3074 * Add blinding to ECDSA and DSA signatures to protect against side channel
3075 attacks discovered by Keegan Ryan (NCC Group).
3079 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3080 now allow empty (zero character) pass phrases.
3084 * Certificate time validation (X509_cmp_time) enforces stricter
3085 compliance with RFC 5280. Fractional seconds and timezone offsets
3086 are no longer allowed.
3090 * Fixed a text canonicalisation bug in CMS
3092 Where a CMS detached signature is used with text content the text goes
3093 through a canonicalisation process first prior to signing or verifying a
3094 signature. This process strips trailing space at the end of lines, converts
3095 line terminators to CRLF and removes additional trailing line terminators
3096 at the end of a file. A bug in the canonicalisation process meant that
3097 some characters, such as form-feed, were incorrectly treated as whitespace
3098 and removed. This is contrary to the specification (RFC5485). This fix
3099 could mean that detached text data signed with an earlier version of
3100 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3101 signed with a fixed OpenSSL may fail to verify with an earlier version of
3102 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3103 and use the "-binary" flag (for the "cms" command line application) or set
3104 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3108 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3110 * Constructed ASN.1 types with a recursive definition could exceed the stack
3112 Constructed ASN.1 types with a recursive definition (such as can be found
3113 in PKCS7) could eventually exceed the stack given malicious input with
3114 excessive recursion. This could result in a Denial Of Service attack. There
3115 are no such structures used within SSL/TLS that come from untrusted sources
3116 so this is considered safe.
3118 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3124 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3126 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3127 effectively reduced to only comparing the least significant bit of each
3128 byte. This allows an attacker to forge messages that would be considered as
3129 authenticated in an amount of tries lower than that guaranteed by the
3130 security claims of the scheme. The module can only be compiled by the
3131 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3133 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3139 * Add a build target 'build_all_generated', to build all generated files
3140 and only that. This can be used to prepare everything that requires
3141 things like perl for a system that lacks perl and then move everything
3142 to that system and do the rest of the build there.
3146 * Backport SSL_OP_NO_RENGOTIATION
3148 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3149 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3150 changes this is no longer possible in 1.1.0. Therefore the new
3151 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3152 1.1.0 to provide equivalent functionality.
3154 Note that if an application built against 1.1.0h headers (or above) is run
3155 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3156 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3160 * Removed the OS390-Unix config target. It relied on a script that doesn't
3165 * rsaz_1024_mul_avx2 overflow bug on x86_64
3167 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3168 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3169 Analysis suggests that attacks against RSA and DSA as a result of this
3170 defect would be very difficult to perform and are not believed likely.
3171 Attacks against DH1024 are considered just feasible, because most of the
3172 work necessary to deduce information about a private key may be performed
3173 offline. The amount of resources required for such an attack would be
3174 significant. However, for an attack on TLS to be meaningful, the server
3175 would have to share the DH1024 private key among multiple clients, which is
3176 no longer an option since CVE-2016-0701.
3178 This only affects processors that support the AVX2 but not ADX extensions
3179 like Intel Haswell (4th generation).
3181 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3182 was originally found via the OSS-Fuzz project.
3187 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3189 * bn_sqrx8x_internal carry bug on x86_64
3191 There is a carry propagating bug in the x86_64 Montgomery squaring
3192 procedure. No EC algorithms are affected. Analysis suggests that attacks
3193 against RSA and DSA as a result of this defect would be very difficult to
3194 perform and are not believed likely. Attacks against DH are considered just
3195 feasible (although very difficult) because most of the work necessary to
3196 deduce information about a private key may be performed offline. The amount
3197 of resources required for such an attack would be very significant and
3198 likely only accessible to a limited number of attackers. An attacker would
3199 additionally need online access to an unpatched system using the target
3200 private key in a scenario with persistent DH parameters and a private
3201 key that is shared between multiple clients.
3203 This only affects processors that support the BMI1, BMI2 and ADX extensions
3204 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3206 This issue was reported to OpenSSL by the OSS-Fuzz project.
3211 * Malformed X.509 IPAddressFamily could cause OOB read
3213 If an X.509 certificate has a malformed IPAddressFamily extension,
3214 OpenSSL could do a one-byte buffer overread. The most likely result
3215 would be an erroneous display of the certificate in text format.
3217 This issue was reported to OpenSSL by the OSS-Fuzz project.
3222 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3224 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3225 platform rather than 'mingw'.
3229 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3230 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3231 which is the minimum version we support.
3235 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3237 * Encrypt-Then-Mac renegotiation crash
3239 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3240 negotiated where it was not in the original handshake (or vice-versa) then
3241 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3242 and servers are affected.
3244 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3249 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3251 * Truncated packet could crash via OOB read
3253 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3254 cipher is being used, then a truncated packet can cause that host to
3255 perform an out-of-bounds read, usually resulting in a crash.
3257 This issue was reported to OpenSSL by Robert Święcki of Google.
3262 * Bad (EC)DHE parameters cause a client crash
3264 If a malicious server supplies bad parameters for a DHE or ECDHE key
3265 exchange then this can result in the client attempting to dereference a
3266 NULL pointer leading to a client crash. This could be exploited in a Denial
3269 This issue was reported to OpenSSL by Guido Vranken.
3274 * BN_mod_exp may produce incorrect results on x86_64
3276 There is a carry propagating bug in the x86_64 Montgomery squaring
3277 procedure. No EC algorithms are affected. Analysis suggests that attacks
3278 against RSA and DSA as a result of this defect would be very difficult to
3279 perform and are not believed likely. Attacks against DH are considered just
3280 feasible (although very difficult) because most of the work necessary to
3281 deduce information about a private key may be performed offline. The amount
3282 of resources required for such an attack would be very significant and
3283 likely only accessible to a limited number of attackers. An attacker would
3284 additionally need online access to an unpatched system using the target
3285 private key in a scenario with persistent DH parameters and a private
3286 key that is shared between multiple clients. For example this can occur by
3287 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3288 similar to CVE-2015-3193 but must be treated as a separate problem.
3290 This issue was reported to OpenSSL by the OSS-Fuzz project.
3295 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3297 * ChaCha20/Poly1305 heap-buffer-overflow
3299 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3300 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3301 crash. This issue is not considered to be exploitable beyond a DoS.
3303 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3308 * CMS Null dereference
3310 Applications parsing invalid CMS structures can crash with a NULL pointer
3311 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3312 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3313 structure callback if an attempt is made to free certain invalid encodings.
3314 Only CHOICE structures using a callback which do not handle NULL value are
3317 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3322 * Montgomery multiplication may produce incorrect results
3324 There is a carry propagating bug in the Broadwell-specific Montgomery
3325 multiplication procedure that handles input lengths divisible by, but
3326 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3327 and DH private keys are impossible. This is because the subroutine in
3328 question is not used in operations with the private key itself and an input
3329 of the attacker's direct choice. Otherwise the bug can manifest itself as
3330 transient authentication and key negotiation failures or reproducible
3331 erroneous outcome of public-key operations with specially crafted input.
3332 Among EC algorithms only Brainpool P-512 curves are affected and one
3333 presumably can attack ECDH key negotiation. Impact was not analyzed in
3334 detail, because pre-requisites for attack are considered unlikely. Namely
3335 multiple clients have to choose the curve in question and the server has to
3336 share the private key among them, neither of which is default behaviour.
3337 Even then only clients that chose the curve will be affected.
3339 This issue was publicly reported as transient failures and was not
3340 initially recognized as a security issue. Thanks to Richard Morgan for
3341 providing reproducible case.
3346 * Removed automatic addition of RPATH in shared libraries and executables,
3347 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3351 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3353 * Fix Use After Free for large message sizes
3355 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3356 message larger than approx 16k is received then the underlying buffer to
3357 store the incoming message is reallocated and moved. Unfortunately a
3358 dangling pointer to the old location is left which results in an attempt to
3359 write to the previously freed location. This is likely to result in a
3360 crash, however it could potentially lead to execution of arbitrary code.
3362 This issue only affects OpenSSL 1.1.0a.
3364 This issue was reported to OpenSSL by Robert Święcki.
3369 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3371 * OCSP Status Request extension unbounded memory growth
3373 A malicious client can send an excessively large OCSP Status Request
3374 extension. If that client continually requests renegotiation, sending a
3375 large OCSP Status Request extension each time, then there will be unbounded
3376 memory growth on the server. This will eventually lead to a Denial Of
3377 Service attack through memory exhaustion. Servers with a default
3378 configuration are vulnerable even if they do not support OCSP. Builds using
3379 the "no-ocsp" build time option are not affected.
3381 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3386 * SSL_peek() hang on empty record
3388 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3389 sends an empty record. This could be exploited by a malicious peer in a
3390 Denial Of Service attack.
3392 This issue was reported to OpenSSL by Alex Gaynor.
3397 * Excessive allocation of memory in tls_get_message_header() and
3398 dtls1_preprocess_fragment()
3400 A (D)TLS message includes 3 bytes for its length in the header for the
3401 message. This would allow for messages up to 16Mb in length. Messages of
3402 this length are excessive and OpenSSL includes a check to ensure that a
3403 peer is sending reasonably sized messages in order to avoid too much memory
3404 being consumed to service a connection. A flaw in the logic of version
3405 1.1.0 means that memory for the message is allocated too early, prior to
3406 the excessive message length check. Due to way memory is allocated in
3407 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3408 to service a connection. This could lead to a Denial of Service through
3409 memory exhaustion. However, the excessive message length check still takes
3410 place, and this would cause the connection to immediately fail. Assuming
3411 that the application calls SSL_free() on the failed connection in a timely
3412 manner then the 21Mb of allocated memory will then be immediately freed
3413 again. Therefore the excessive memory allocation will be transitory in
3414 nature. This then means that there is only a security impact if:
3416 1) The application does not call SSL_free() in a timely manner in the event
3417 that the connection fails
3419 2) The application is working in a constrained environment where there is
3420 very little free memory
3422 3) The attacker initiates multiple connection attempts such that there are
3423 multiple connections in a state where memory has been allocated for the
3424 connection; SSL_free() has not yet been called; and there is insufficient
3425 memory to service the multiple requests.
3427 Except in the instance of (1) above any Denial Of Service is likely to be
3428 transitory because as soon as the connection fails the memory is
3429 subsequently freed again in the SSL_free() call. However there is an
3430 increased risk during this period of application crashes due to the lack of
3431 memory - which would then mean a more serious Denial of Service.
3433 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3434 (CVE-2016-6307 and CVE-2016-6308)
3438 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3439 had to be removed. Primary reason is that vendor assembler can't
3440 assemble our modules with -KPIC flag. As result it, assembly
3441 support, was not even available as option. But its lack means
3442 lack of side-channel resistant code, which is incompatible with
3443 security by todays standards. Fortunately gcc is readily available
3444 prepackaged option, which we firmly point at...
3448 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3450 * Windows command-line tool supports UTF-8 opt-in option for arguments
3451 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3452 (to any value) allows Windows user to access PKCS#12 file generated
3453 with Windows CryptoAPI and protected with non-ASCII password, as well
3454 as files generated under UTF-8 locale on Linux also protected with
3459 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3460 have been disabled by default and removed from DEFAULT, just like RC4.
3461 See the RC4 item below to re-enable both.
3465 * The method for finding the storage location for the Windows RAND seed file
3466 has changed. First we check %RANDFILE%. If that is not set then we check
3467 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3468 all else fails we fall back to C:\.
3472 * The EVP_EncryptUpdate() function has had its return type changed from void
3473 to int. A return of 0 indicates and error while a return of 1 indicates
3478 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3479 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3480 off the constant time implementation for RSA, DSA and DH have been made
3481 no-ops and deprecated.
3485 * Windows RAND implementation was simplified to only get entropy by
3486 calling CryptGenRandom(). Various other RAND-related tickets
3489 *Joseph Wylie Yandle, Rich Salz*
3491 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3492 and `OPENSSL_LH_`, respectively. The old names are available
3493 with API compatibility. They new names are now completely documented.
3497 * Unify TYPE_up_ref(obj) methods signature.
3498 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3499 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3500 int (instead of void) like all others TYPE_up_ref() methods.
3501 So now these methods also check the return value of CRYPTO_atomic_add(),
3502 and the validity of object reference counter.
3504 *fdasilvayy@gmail.com*
3506 * With Windows Visual Studio builds, the .pdb files are installed
3507 alongside the installed libraries and executables. For a static
3508 library installation, ossl_static.pdb is the associate compiler
3509 generated .pdb file to be used when linking programs.
3513 * Remove openssl.spec. Packaging files belong with the packagers.
3517 * Automatic Darwin/OSX configuration has had a refresh, it will now
3518 recognise x86_64 architectures automatically. You can still decide
3519 to build for a different bitness with the environment variable
3520 KERNEL_BITS (can be 32 or 64), for example:
3522 KERNEL_BITS=32 ./config
3526 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3527 256 bit AES and HMAC with SHA256.
3531 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3535 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3539 * To enable users to have their own config files and build file templates,
3540 Configure looks in the directory indicated by the environment variable
3541 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3542 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3543 name and is used as is.
3547 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3548 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3549 X509_CERT_FILE_CTX was removed.
3553 * "shared" builds are now the default. To create only static libraries use
3554 the "no-shared" Configure option.
3558 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3559 All of these option have not worked for some while and are fundamental
3564 * Make various cleanup routines no-ops and mark them as deprecated. Most
3565 global cleanup functions are no longer required because they are handled
3566 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3567 Explicitly de-initing can cause problems (e.g. where a library that uses
3568 OpenSSL de-inits, but an application is still using it). The affected
3569 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3570 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3571 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3572 COMP_zlib_cleanup().
3576 * --strict-warnings no longer enables runtime debugging options
3577 such as REF_DEBUG. Instead, debug options are automatically
3578 enabled with '--debug' builds.
3580 *Andy Polyakov, Emilia Käsper*
3582 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3583 have been moved out of the public header files. New functions for managing
3584 these have been added.
3588 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3589 objects have been moved out of the public header files. New
3590 functions for managing these have been added.
3594 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3595 have been moved out of the public header files. New functions for managing
3596 these have been added.
3600 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3601 moved out of the public header files. New functions for managing these
3606 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3610 * Removed the mk1mf build scripts.
3614 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3615 it is always safe to #include a header now.
3619 * Removed the aged BC-32 config and all its supporting scripts
3623 * Removed support for Ultrix, Netware, and OS/2.
3627 * Add support for HKDF.
3629 *Alessandro Ghedini*
3631 * Add support for blake2b and blake2s
3635 * Added support for "pipelining". Ciphers that have the
3636 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3637 encryptions/decryptions simultaneously. There are currently no built-in
3638 ciphers with this property but the expectation is that engines will be able
3639 to offer it to significantly improve throughput. Support has been extended
3640 into libssl so that multiple records for a single connection can be
3641 processed in one go (for >=TLS 1.1).
3645 * Added the AFALG engine. This is an async capable engine which is able to
3646 offload work to the Linux kernel. In this initial version it only supports
3647 AES128-CBC. The kernel must be version 4.1.0 or greater.
3651 * OpenSSL now uses a new threading API. It is no longer necessary to
3652 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3653 are two supported threading models: pthreads and windows threads. It is
3654 also possible to configure OpenSSL at compile time for "no-threads". The
3655 old threading API should no longer be used. The functions have been
3656 replaced with "no-op" compatibility macros.
3658 *Alessandro Ghedini, Matt Caswell*
3660 * Modify behavior of ALPN to invoke callback after SNI/servername
3661 callback, such that updates to the SSL_CTX affect ALPN.
3665 * Add SSL_CIPHER queries for authentication and key-exchange.
3669 * Changes to the DEFAULT cipherlist:
3670 - Prefer (EC)DHE handshakes over plain RSA.
3671 - Prefer AEAD ciphers over legacy ciphers.
3672 - Prefer ECDSA over RSA when both certificates are available.
3673 - Prefer TLSv1.2 ciphers/PRF.
3674 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3679 * Change the ECC default curve list to be this, in order: x25519,
3680 secp256r1, secp521r1, secp384r1.
3684 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3685 disabled by default. They can be re-enabled using the
3686 enable-weak-ssl-ciphers option to Configure.
3690 * If the server has ALPN configured, but supports no protocols that the
3691 client advertises, send a fatal "no_application_protocol" alert.
3692 This behaviour is SHALL in RFC 7301, though it isn't universally
3693 implemented by other servers.
3697 * Add X25519 support.
3698 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3699 for public and private key encoding using the format documented in
3700 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3701 key generation and key derivation.
3703 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3708 * Deprecate SRP_VBASE_get_by_user.
3709 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3710 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3711 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3712 seed, even if the seed is configured.
3714 Users should use SRP_VBASE_get1_by_user instead. Note that in
3715 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3716 also that even though configuring the SRP seed attempts to hide
3717 invalid usernames by continuing the handshake with fake
3718 credentials, this behaviour is not constant time and no strong
3719 guarantees are made that the handshake is indistinguishable from
3720 that of a valid user.
3724 * Configuration change; it's now possible to build dynamic engines
3725 without having to build shared libraries and vice versa. This
3726 only applies to the engines in `engines/`, those in `crypto/engine/`
3727 will always be built into libcrypto (i.e. "static").
3729 Building dynamic engines is enabled by default; to disable, use
3730 the configuration option "disable-dynamic-engine".
3732 The only requirements for building dynamic engines are the
3733 presence of the DSO module and building with position independent
3734 code, so they will also automatically be disabled if configuring
3735 with "disable-dso" or "disable-pic".
3737 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3738 are also taken away from openssl/opensslconf.h, as they are
3743 * Configuration change; if there is a known flag to compile
3744 position independent code, it will always be applied on the
3745 libcrypto and libssl object files, and never on the application
3746 object files. This means other libraries that use routines from
3747 libcrypto / libssl can be made into shared libraries regardless
3748 of how OpenSSL was configured.
3750 If this isn't desirable, the configuration options "disable-pic"
3751 or "no-pic" can be used to disable the use of PIC. This will
3752 also disable building shared libraries and dynamic engines.
3756 * Removed JPAKE code. It was experimental and has no wide use.
3760 * The INSTALL_PREFIX Makefile variable has been renamed to
3761 DESTDIR. That makes for less confusion on what this variable
3762 is for. Also, the configuration option --install_prefix is
3767 * Heartbeat for TLS has been removed and is disabled by default
3768 for DTLS; configure with enable-heartbeats. Code that uses the
3769 old #define's might need to be updated.
3771 *Emilia Käsper, Rich Salz*
3773 * Rename REF_CHECK to REF_DEBUG.
3777 * New "unified" build system
3779 The "unified" build system is aimed to be a common system for all
3780 platforms we support. With it comes new support for VMS.
3782 This system builds supports building in a different directory tree
3783 than the source tree. It produces one Makefile (for unix family
3784 or lookalikes), or one descrip.mms (for VMS).
3786 The source of information to make the Makefile / descrip.mms is
3787 small files called 'build.info', holding the necessary
3788 information for each directory with source to compile, and a
3789 template in Configurations, like unix-Makefile.tmpl or
3792 With this change, the library names were also renamed on Windows
3793 and on VMS. They now have names that are closer to the standard
3794 on Unix, and include the major version number, and in certain
3795 cases, the architecture they are built for. See "Notes on shared
3796 libraries" in INSTALL.
3798 We rely heavily on the perl module Text::Template.
3802 * Added support for auto-initialisation and de-initialisation of the library.
3803 OpenSSL no longer requires explicit init or deinit routines to be called,
3804 except in certain circumstances. See the OPENSSL_init_crypto() and
3805 OPENSSL_init_ssl() man pages for further information.
3809 * The arguments to the DTLSv1_listen function have changed. Specifically the
3810 "peer" argument is now expected to be a BIO_ADDR object.
3812 * Rewrite of BIO networking library. The BIO library lacked consistent
3813 support of IPv6, and adding it required some more extensive
3814 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3815 which hold all types of addresses and chains of address information.
3816 It also introduces a new API, with functions like BIO_socket,
3817 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3818 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3819 have been adapted accordingly.
3823 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3828 * CRIME protection: disable compression by default, even if OpenSSL is
3829 compiled with zlib enabled. Applications can still enable compression
3830 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3831 using the SSL_CONF library to configure compression.
3835 * The signature of the session callback configured with
3836 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3837 was explicitly marked as `const unsigned char*` instead of
3842 * Always DPURIFY. Remove the use of uninitialized memory in the
3843 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3847 * Removed many obsolete configuration items, including
3848 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3849 MD2_CHAR, MD2_INT, MD2_LONG
3851 IDEA_SHORT, IDEA_LONG
3852 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3854 *Rich Salz, with advice from Andy Polyakov*
3856 * Many BN internals have been moved to an internal header file.
3858 *Rich Salz with help from Andy Polyakov*
3860 * Configuration and writing out the results from it has changed.
3861 Files such as Makefile include/openssl/opensslconf.h and are now
3862 produced through general templates, such as Makefile.in and
3863 crypto/opensslconf.h.in and some help from the perl module
3866 Also, the center of configuration information is no longer
3867 Makefile. Instead, Configure produces a perl module in
3868 configdata.pm which holds most of the config data (in the hash
3869 table %config), the target data that comes from the target
3870 configuration in one of the `Configurations/*.conf` files (in
3875 * To clarify their intended purposes, the Configure options
3876 --prefix and --openssldir change their semantics, and become more
3877 straightforward and less interdependent.
3879 --prefix shall be used exclusively to give the location INSTALLTOP
3880 where programs, scripts, libraries, include files and manuals are
3881 going to be installed. The default is now /usr/local.
3883 --openssldir shall be used exclusively to give the default
3884 location OPENSSLDIR where certificates, private keys, CRLs are
3885 managed. This is also where the default openssl.cnf gets
3887 If the directory given with this option is a relative path, the
3888 values of both the --prefix value and the --openssldir value will
3889 be combined to become OPENSSLDIR.
3890 The default for --openssldir is INSTALLTOP/ssl.
3892 Anyone who uses --openssldir to specify where OpenSSL is to be
3893 installed MUST change to use --prefix instead.
3897 * The GOST engine was out of date and therefore it has been removed. An up
3898 to date GOST engine is now being maintained in an external repository.
3899 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3900 support for GOST ciphersuites (these are only activated if a GOST engine
3905 * EGD is no longer supported by default; use enable-egd when
3908 *Ben Kaduk and Rich Salz*
3910 * The distribution now has Makefile.in files, which are used to
3911 create Makefile's when Configure is run. *Configure must be run
3912 before trying to build now.*
3916 * The return value for SSL_CIPHER_description() for error conditions
3921 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3923 Obtaining and performing DNSSEC validation of TLSA records is
3924 the application's responsibility. The application provides
3925 the TLSA records of its choice to OpenSSL, and these are then
3926 used to authenticate the peer.
3928 The TLSA records need not even come from DNS. They can, for
3929 example, be used to implement local end-entity certificate or
3930 trust-anchor "pinning", where the "pin" data takes the form
3931 of TLSA records, which can augment or replace verification
3932 based on the usual WebPKI public certification authorities.
3936 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3937 continues to support deprecated interfaces in default builds.
3938 However, applications are strongly advised to compile their
3939 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3940 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3941 or the 1.1.0 releases.
3943 In environments in which all applications have been ported to
3944 not use any deprecated interfaces OpenSSL's Configure script
3945 should be used with the --api=1.1.0 option to entirely remove
3946 support for the deprecated features from the library and
3947 unconditionally disable them in the installed headers.
3948 Essentially the same effect can be achieved with the "no-deprecated"
3949 argument to Configure, except that this will always restrict
3950 the build to just the latest API, rather than a fixed API
3953 As applications are ported to future revisions of the API,
3954 they should update their compile-time OPENSSL_API_COMPAT define
3955 accordingly, but in most cases should be able to continue to
3956 compile with later releases.
3958 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3959 0x10000000L and 0x00908000L, respectively. However those
3960 versions did not support the OPENSSL_API_COMPAT feature, and
3961 so applications are not typically tested for explicit support
3962 of just the undeprecated features of either release.
3966 * Add support for setting the minimum and maximum supported protocol.
3967 It can bet set via the SSL_set_min_proto_version() and
3968 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3969 MaxProtocol. It's recommended to use the new APIs to disable
3970 protocols instead of disabling individual protocols using
3971 SSL_set_options() or SSL_CONF's Protocol. This change also
3972 removes support for disabling TLS 1.2 in the OpenSSL TLS
3973 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3977 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3981 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3982 and integrates ECDSA and ECDH functionality into EC. Implementations can
3983 now redirect key generation and no longer need to convert to or from
3986 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3987 include the ec.h header file instead.
3991 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3992 ciphers who are no longer supported and drops support the ephemeral RSA key
3993 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3997 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3998 opaque. For HMAC_CTX, the following constructors and destructors
4001 HMAC_CTX *HMAC_CTX_new(void);
4002 void HMAC_CTX_free(HMAC_CTX *ctx);
4004 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4005 destroy such methods has been added. See EVP_MD_meth_new(3) and
4006 EVP_CIPHER_meth_new(3) for documentation.
4009 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4010 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4011 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4012 an already created structure.
4013 2) For consistency with the majority of our object creators and
4014 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4015 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4016 for deprecated builds.
4020 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4021 cryptographic operations to be performed asynchronously as long as an
4022 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4023 further details. Libssl has also had this capability integrated with the
4024 introduction of the new mode SSL_MODE_ASYNC and associated error
4025 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4026 pages. This work was developed in partnership with Intel Corp.
4030 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4031 always enabled now. If you want to disable the support you should
4032 exclude it using the list of supported ciphers. This also means that the
4033 "-no_ecdhe" option has been removed from s_server.
4037 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4038 SSL_{CTX_}set1_curves() which can set a list.
4042 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4043 curve you want to support using SSL_{CTX_}set1_curves().
4047 * State machine rewrite. The state machine code has been significantly
4048 refactored in order to remove much duplication of code and solve issues
4049 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4050 further details). This change does have some associated API changes.
4051 Notably the SSL_state() function has been removed and replaced by
4052 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4053 SSL_set_state() has been removed altogether. The previous handshake states
4054 defined in ssl.h and ssl3.h have also been removed.
4058 * All instances of the string "ssleay" in the public API were replaced
4059 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4060 Some error codes related to internal RSA_eay API's were renamed.
4064 * The demo files in crypto/threads were moved to demo/threads.
4068 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4071 *Matt Caswell, Rich Salz*
4073 * New ASN.1 embed macro.
4075 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4076 structure is not allocated: it is part of the parent. That is instead of
4084 This reduces memory fragmentation and make it impossible to accidentally
4085 set a mandatory field to NULL.
4087 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4088 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4089 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4094 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4098 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4099 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4100 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4101 DES and RC4 ciphersuites.
4105 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4106 This changes the decoding behaviour for some invalid messages,
4107 though the change is mostly in the more lenient direction, and
4108 legacy behaviour is preserved as much as possible.
4112 * Fix no-stdio build.
4113 *David Woodhouse <David.Woodhouse@intel.com> and also*
4114 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4116 * New testing framework
4117 The testing framework has been largely rewritten and is now using
4118 perl and the perl modules Test::Harness and an extended variant of
4119 Test::More called OpenSSL::Test to do its work. All test scripts in
4120 test/ have been rewritten into test recipes, and all direct calls to
4121 executables in test/Makefile have become individual recipes using the
4122 simplified testing OpenSSL::Test::Simple.
4124 For documentation on our testing modules, do:
4126 perldoc test/testlib/OpenSSL/Test/Simple.pm
4127 perldoc test/testlib/OpenSSL/Test.pm
4131 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4132 are used; the latter aborts on memory leaks (usually checked on exit).
4133 Some undocumented "set malloc, etc., hooks" functions were removed
4134 and others were changed. All are now documented.
4138 * In DSA_generate_parameters_ex, if the provided seed is too short,
4141 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4143 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4144 from RFC4279, RFC4785, RFC5487, RFC5489.
4146 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4147 original RSA_PSK patch.
4151 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4152 era flag was never set throughout the codebase (only read). Also removed
4153 SSL3_FLAGS_POP_BUFFER which was only used if
4154 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4158 * Changed the default name options in the "ca", "crl", "req" and "x509"
4159 to be "oneline" instead of "compat".
4163 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4164 not aware of clients that still exhibit this bug, and the workaround
4165 hasn't been working properly for a while.
4169 * The return type of BIO_number_read() and BIO_number_written() as well as
4170 the corresponding num_read and num_write members in the BIO structure has
4171 changed from unsigned long to uint64_t. On platforms where an unsigned
4172 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4177 * Given the pervasive nature of TLS extensions it is inadvisable to run
4178 OpenSSL without support for them. It also means that maintaining
4179 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4180 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4184 * Removed support for the two export grade static DH ciphersuites
4185 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4186 were newly added (along with a number of other static DH ciphersuites) to
4187 1.0.2. However the two export ones have *never* worked since they were
4188 introduced. It seems strange in any case to be adding new export
4189 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4193 * Version negotiation has been rewritten. In particular SSLv23_method(),
4194 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4195 and turned into macros which simply call the new preferred function names
4196 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4197 should use the new names instead. Also as part of this change the ssl23.h
4198 header file has been removed.
4202 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4203 code and the associated standard is no longer considered fit-for-purpose.
4207 * RT2547 was closed. When generating a private key, try to make the
4208 output file readable only by the owner. This behavior change might
4209 be noticeable when interacting with other software.
4211 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4216 * Added HTTP GET support to the ocsp command.
4220 * Changed default digest for the dgst and enc commands from MD5 to
4225 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4229 * Added support for TLS extended master secret from
4230 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4231 initial patch which was a great help during development.
4235 * All libssl internal structures have been removed from the public header
4236 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4237 now redundant). Users should not attempt to access internal structures
4238 directly. Instead they should use the provided API functions.
4242 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4243 Access to deprecated functions can be re-enabled by running config with
4244 "enable-deprecated". In addition applications wishing to use deprecated
4245 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4246 will, by default, disable some transitive includes that previously existed
4247 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4251 * Added support for OCB mode. OpenSSL has been granted a patent license
4252 compatible with the OpenSSL license for use of OCB. Details are available
4253 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4254 for OCB can be removed by calling config with no-ocb.
4258 * SSLv2 support has been removed. It still supports receiving a SSLv2
4259 compatible client hello.
4263 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4264 done while fixing the error code for the key-too-small case.
4266 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4268 * CA.sh has been removed; use CA.pl instead.
4272 * Removed old DES API.
4276 * Remove various unsupported platforms:
4282 Sinix/ReliantUNIX RM400
4287 16-bit platforms such as WIN16
4291 * Clean up OPENSSL_NO_xxx #define's
4292 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4293 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4294 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4295 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4296 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4297 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4298 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4299 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4300 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4301 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4305 * Cleaned up dead code
4306 Remove all but one '#ifdef undef' which is to be looked at.
4310 * Clean up calling of xxx_free routines.
4311 Just like free(), fix most of the xxx_free routines to accept
4312 NULL. Remove the non-null checks from callers. Save much code.
4316 * Add secure heap for storage of private keys (when possible).
4317 Add BIO_s_secmem(), CBIGNUM, etc.
4318 Contributed by Akamai Technologies under our Corporate CLA.
4322 * Experimental support for a new, fast, unbiased prime candidate generator,
4323 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4325 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4327 * New output format NSS in the sess_id command line tool. This allows
4328 exporting the session id and the master key in NSS keylog format.
4330 *Martin Kaiser <martin@kaiser.cx>*
4332 * Harmonize version and its documentation. -f flag is used to display
4335 *mancha <mancha1@zoho.com>*
4337 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4338 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4340 *mancha <mancha1@zoho.com>*
4342 * Fix some double frees. These are not thought to be exploitable.
4344 *mancha <mancha1@zoho.com>*
4346 * A missing bounds check in the handling of the TLS heartbeat extension
4347 can be used to reveal up to 64k of memory to a connected client or
4350 Thanks for Neel Mehta of Google Security for discovering this bug and to
4351 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4352 preparing the fix ([CVE-2014-0160])
4354 *Adam Langley, Bodo Moeller*
4356 * Fix for the attack described in the paper "Recovering OpenSSL
4357 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4358 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4359 <http://eprint.iacr.org/2014/140>
4361 Thanks to Yuval Yarom and Naomi Benger for discovering this
4362 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4364 *Yuval Yarom and Naomi Benger*
4366 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4367 this fixes a limitation in previous versions of OpenSSL.
4371 * Experimental encrypt-then-mac support.
4373 Experimental support for encrypt then mac from
4374 draft-gutmann-tls-encrypt-then-mac-02.txt
4376 To enable it set the appropriate extension number (0x42 for the test
4377 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4379 For non-compliant peers (i.e. just about everything) this should have no
4382 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4386 * Add EVP support for key wrapping algorithms, to avoid problems with
4387 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4388 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4389 algorithms and include tests cases.
4393 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4398 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4399 MGF1 digest and OAEP label.
4403 * Make openssl verify return errors.
4405 *Chris Palmer <palmer@google.com> and Ben Laurie*
4407 * New function ASN1_TIME_diff to calculate the difference between two
4408 ASN1_TIME structures or one structure and the current time.
4412 * Update fips_test_suite to support multiple command line options. New
4413 test to induce all self test errors in sequence and check expected
4418 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4419 sign or verify all in one operation.
4423 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4424 test programs and fips_test_suite. Includes functionality to parse
4425 the minimal script output of fipsalgest.pl directly.
4429 * Add authorisation parameter to FIPS_module_mode_set().
4433 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4437 * Use separate DRBG fields for internal and external flags. New function
4438 FIPS_drbg_health_check() to perform on demand health checking. Add
4439 generation tests to fips_test_suite with reduced health check interval to
4440 demonstrate periodic health checking. Add "nodh" option to
4441 fips_test_suite to skip very slow DH test.
4445 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4450 * More extensive health check for DRBG checking many more failure modes.
4451 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4452 combination: call this in fips_test_suite.
4456 * Add support for canonical generation of DSA parameter 'g'. See
4459 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4460 POST to handle HMAC cases.
4464 * Add functions FIPS_module_version() and FIPS_module_version_text()
4465 to return numerical and string versions of the FIPS module number.
4469 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4470 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4471 outside the validated module in the FIPS capable OpenSSL.
4475 * Minor change to DRBG entropy callback semantics. In some cases
4476 there is no multiple of the block length between min_len and
4477 max_len. Allow the callback to return more than max_len bytes
4478 of entropy but discard any extra: it is the callback's responsibility
4479 to ensure that the extra data discarded does not impact the
4480 requested amount of entropy.
4484 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4485 information in FIPS186-3, SP800-57 and SP800-131A.
4489 * CCM support via EVP. Interface is very similar to GCM case except we
4490 must supply all data in one chunk (i.e. no update, final) and the
4491 message length must be supplied if AAD is used. Add algorithm test
4496 * Initial version of POST overhaul. Add POST callback to allow the status
4497 of POST to be monitored and/or failures induced. Modify fips_test_suite
4498 to use callback. Always run all selftests even if one fails.
4502 * XTS support including algorithm test driver in the fips_gcmtest program.
4503 Note: this does increase the maximum key length from 32 to 64 bytes but
4504 there should be no binary compatibility issues as existing applications
4505 will never use XTS mode.
4509 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4510 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4511 performs algorithm blocking for unapproved PRNG types. Also do not
4512 set PRNG type in FIPS_mode_set(): leave this to the application.
4513 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4514 the standard OpenSSL PRNG: set additional data to a date time vector.
4518 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4519 This shouldn't present any incompatibility problems because applications
4520 shouldn't be using these directly and any that are will need to rethink
4521 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4525 * Extensive self tests and health checking required by SP800-90 DRBG.
4526 Remove strength parameter from FIPS_drbg_instantiate and always
4527 instantiate at maximum supported strength.
4531 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4535 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4539 * New function DH_compute_key_padded() to compute a DH key and pad with
4540 leading zeroes if needed: this complies with SP800-56A et al.
4544 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4545 anything, incomplete, subject to change and largely untested at present.
4549 * Modify fipscanisteronly build option to only build the necessary object
4550 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4554 * Add experimental option FIPSSYMS to give all symbols in
4555 fipscanister.o and FIPS or fips prefix. This will avoid
4556 conflicts with future versions of OpenSSL. Add perl script
4557 util/fipsas.pl to preprocess assembly language source files
4558 and rename any affected symbols.
4562 * Add selftest checks and algorithm block of non-fips algorithms in
4563 FIPS mode. Remove DES2 from selftests.
4567 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4568 return internal method without any ENGINE dependencies. Add new
4569 tiny fips sign and verify functions.
4573 * New build option no-ec2m to disable characteristic 2 code.
4577 * New build option "fipscanisteronly". This only builds fipscanister.o
4578 and (currently) associated fips utilities. Uses the file Makefile.fips
4579 instead of Makefile.org as the prototype.
4583 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4584 Update fips_gcmtest to use IV generator.
4588 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4589 setting output buffer to NULL. The `*Final` function must be
4590 called although it will not retrieve any additional data. The tag
4591 can be set or retrieved with a ctrl. The IV length is by default 12
4592 bytes (96 bits) but can be set to an alternative value. If the IV
4593 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4598 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4599 underlying do_cipher function handles all cipher semantics itself
4600 including padding and finalisation. This is useful if (for example)
4601 an ENGINE cipher handles block padding itself. The behaviour of
4602 do_cipher is subtly changed if this flag is set: the return value
4603 is the number of characters written to the output buffer (zero is
4604 no longer an error code) or a negative error code. Also if the
4605 input buffer is NULL and length 0 finalisation should be performed.
4609 * If a candidate issuer certificate is already part of the constructed
4610 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4614 * Improve forward-security support: add functions
4616 void SSL_CTX_set_not_resumable_session_callback(
4617 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4618 void SSL_set_not_resumable_session_callback(
4619 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4621 for use by SSL/TLS servers; the callback function will be called whenever a
4622 new session is created, and gets to decide whether the session may be
4623 cached to make it resumable (return 0) or not (return 1). (As by the
4624 SSL/TLS protocol specifications, the session_id sent by the server will be
4625 empty to indicate that the session is not resumable; also, the server will
4626 not generate RFC 4507 (RFC 5077) session tickets.)
4628 A simple reasonable callback implementation is to return is_forward_secure.
4629 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4630 by the SSL/TLS server library, indicating whether it can provide forward
4633 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4635 * New -verify_name option in command line utilities to set verification
4640 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4641 Add CMAC pkey methods.
4645 * Experimental renegotiation in s_server -www mode. If the client
4646 browses /reneg connection is renegotiated. If /renegcert it is
4647 renegotiated requesting a certificate.
4651 * Add an "external" session cache for debugging purposes to s_server. This
4652 should help trace issues which normally are only apparent in deployed
4653 multi-process servers.
4657 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4658 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4659 BIO_set_cipher() and some obscure PEM functions were changed so they
4660 can now return an error. The RAND changes required a change to the
4661 RAND_METHOD structure.
4665 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4666 a gcc attribute to warn if the result of a function is ignored. This
4667 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4668 whose return value is often ignored.
4672 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4673 These allow SCTs (signed certificate timestamps) to be requested and
4674 validated when establishing a connection.
4676 *Rob Percival <robpercival@google.com>*
4681 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4683 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4684 used even when parsing explicit parameters, when loading a encoded key
4685 or calling `EC_GROUP_new_from_ecpkparameters()`/
4686 `EC_GROUP_new_from_ecparameters()`.
4687 This prevents bypass of security hardening and performance gains,
4688 especially for curves with specialized EC_METHODs.
4689 By default, if a key encoded with explicit parameters is loaded and later
4690 encoded, the output is still encoded with explicit parameters, even if
4691 internally a "named" EC_GROUP is used for computation.
4695 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4696 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4697 NULL. After this change, only the cofactor parameter can be NULL. It also
4698 does some minimal sanity checks on the passed order.
4703 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4704 An attack is simple, if the first CMS_recipientInfo is valid but the
4705 second CMS_recipientInfo is chosen ciphertext. If the second
4706 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4707 encryption key will be replaced by garbage, and the message cannot be
4708 decoded, but if the RSA decryption fails, the correct encryption key is
4709 used and the recipient will not notice the attack.
4710 As a work around for this potential attack the length of the decrypted
4711 key must be equal to the cipher default key length, in case the
4712 certifiate is not given and all recipientInfo are tried out.
4713 The old behaviour can be re-enabled in the CMS code by setting the
4714 CMS_DEBUG_DECRYPT flag.
4719 * Document issue with installation paths in diverse Windows builds
4721 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4722 binaries and run-time config file.
4727 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4729 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4730 This changes the size when using the `genpkey` command when no size is given.
4731 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4732 generation commands to use 2048 bits by default.
4736 * Add FIPS support for Android Arm 64-bit
4738 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4739 Module in Version 2.0.10. For some reason, the corresponding target
4740 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4741 built with FIPS support on Android Arm 64-bit. This omission has been
4744 *Matthias St. Pierre*
4746 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4748 * 0-byte record padding oracle
4750 If an application encounters a fatal protocol error and then calls
4751 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4752 then OpenSSL can respond differently to the calling application if a 0 byte
4753 record is received with invalid padding compared to if a 0 byte record is
4754 received with an invalid MAC. If the application then behaves differently
4755 based on that in a way that is detectable to the remote peer, then this
4756 amounts to a padding oracle that could be used to decrypt data.
4758 In order for this to be exploitable "non-stitched" ciphersuites must be in
4759 use. Stitched ciphersuites are optimised implementations of certain
4760 commonly used ciphersuites. Also the application must call SSL_shutdown()
4761 twice even if a protocol error has occurred (applications should not do
4762 this but some do anyway).
4764 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4765 Aviram, with additional investigation by Steven Collison and Andrew
4766 Hourselt. It was reported to OpenSSL on 10th December 2018.
4771 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4775 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4777 * Microarchitecture timing vulnerability in ECC scalar multiplication
4779 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4780 shown to be vulnerable to a microarchitecture timing side channel attack.
4781 An attacker with sufficient access to mount local timing attacks during
4782 ECDSA signature generation could recover the private key.
4784 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4785 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4791 * Timing vulnerability in DSA signature generation
4793 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4794 timing side channel attack. An attacker could use variations in the signing
4795 algorithm to recover the private key.
4797 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4802 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4803 Module, accidentally introduced while backporting security fixes from the
4804 development branch and hindering the use of ECC in FIPS mode.
4808 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4810 * Client DoS due to large DH parameter
4812 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4813 malicious server can send a very large prime value to the client. This will
4814 cause the client to spend an unreasonably long period of time generating a
4815 key for this prime resulting in a hang until the client has finished. This
4816 could be exploited in a Denial Of Service attack.
4818 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4823 * Cache timing vulnerability in RSA Key Generation
4825 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4826 a cache timing side channel attack. An attacker with sufficient access to
4827 mount cache timing attacks during the RSA key generation process could
4828 recover the private key.
4830 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4831 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4836 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4837 parameter is no longer accepted, as it leads to a corrupt table. NULL
4838 pem_str is reserved for alias entries only.
4842 * Revert blinding in ECDSA sign and instead make problematic addition
4843 length-invariant. Switch even to fixed-length Montgomery multiplication.
4847 * Change generating and checking of primes so that the error rate of not
4848 being prime depends on the intended use based on the size of the input.
4849 For larger primes this will result in more rounds of Miller-Rabin.
4850 The maximal error rate for primes with more than 1080 bits is lowered
4853 *Kurt Roeckx, Annie Yousar*
4855 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4859 * Add blinding to ECDSA and DSA signatures to protect against side channel
4860 attacks discovered by Keegan Ryan (NCC Group).
4864 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4865 now allow empty (zero character) pass phrases.
4869 * Certificate time validation (X509_cmp_time) enforces stricter
4870 compliance with RFC 5280. Fractional seconds and timezone offsets
4871 are no longer allowed.
4875 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4877 * Constructed ASN.1 types with a recursive definition could exceed the stack
4879 Constructed ASN.1 types with a recursive definition (such as can be found
4880 in PKCS7) could eventually exceed the stack given malicious input with
4881 excessive recursion. This could result in a Denial Of Service attack. There
4882 are no such structures used within SSL/TLS that come from untrusted sources
4883 so this is considered safe.
4885 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4891 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4893 * Read/write after SSL object in error state
4895 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4896 mechanism. The intent was that if a fatal error occurred during a handshake
4897 then OpenSSL would move into the error state and would immediately fail if
4898 you attempted to continue the handshake. This works as designed for the
4899 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4900 SSL_connect()), however due to a bug it does not work correctly if
4901 SSL_read() or SSL_write() is called directly. In that scenario, if the
4902 handshake fails then a fatal error will be returned in the initial function
4903 call. If SSL_read()/SSL_write() is subsequently called by the application
4904 for the same SSL object then it will succeed and the data is passed without
4905 being decrypted/encrypted directly from the SSL/TLS record layer.
4907 In order to exploit this issue an application bug would have to be present
4908 that resulted in a call to SSL_read()/SSL_write() being issued after having
4909 already received a fatal error.
4911 This issue was reported to OpenSSL by David Benjamin (Google).
4916 * rsaz_1024_mul_avx2 overflow bug on x86_64
4918 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4919 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4920 Analysis suggests that attacks against RSA and DSA as a result of this
4921 defect would be very difficult to perform and are not believed likely.
4922 Attacks against DH1024 are considered just feasible, because most of the
4923 work necessary to deduce information about a private key may be performed
4924 offline. The amount of resources required for such an attack would be
4925 significant. However, for an attack on TLS to be meaningful, the server
4926 would have to share the DH1024 private key among multiple clients, which is
4927 no longer an option since CVE-2016-0701.
4929 This only affects processors that support the AVX2 but not ADX extensions
4930 like Intel Haswell (4th generation).
4932 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4933 was originally found via the OSS-Fuzz project.
4938 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4940 * bn_sqrx8x_internal carry bug on x86_64
4942 There is a carry propagating bug in the x86_64 Montgomery squaring
4943 procedure. No EC algorithms are affected. Analysis suggests that attacks
4944 against RSA and DSA as a result of this defect would be very difficult to
4945 perform and are not believed likely. Attacks against DH are considered just
4946 feasible (although very difficult) because most of the work necessary to
4947 deduce information about a private key may be performed offline. The amount
4948 of resources required for such an attack would be very significant and
4949 likely only accessible to a limited number of attackers. An attacker would
4950 additionally need online access to an unpatched system using the target
4951 private key in a scenario with persistent DH parameters and a private
4952 key that is shared between multiple clients.
4954 This only affects processors that support the BMI1, BMI2 and ADX extensions
4955 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4957 This issue was reported to OpenSSL by the OSS-Fuzz project.
4962 * Malformed X.509 IPAddressFamily could cause OOB read
4964 If an X.509 certificate has a malformed IPAddressFamily extension,
4965 OpenSSL could do a one-byte buffer overread. The most likely result
4966 would be an erroneous display of the certificate in text format.
4968 This issue was reported to OpenSSL by the OSS-Fuzz project.
4972 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4974 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4975 platform rather than 'mingw'.
4979 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4981 * Truncated packet could crash via OOB read
4983 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4984 cipher is being used, then a truncated packet can cause that host to
4985 perform an out-of-bounds read, usually resulting in a crash.
4987 This issue was reported to OpenSSL by Robert Święcki of Google.
4992 * BN_mod_exp may produce incorrect results on x86_64
4994 There is a carry propagating bug in the x86_64 Montgomery squaring
4995 procedure. No EC algorithms are affected. Analysis suggests that attacks
4996 against RSA and DSA as a result of this defect would be very difficult to
4997 perform and are not believed likely. Attacks against DH are considered just
4998 feasible (although very difficult) because most of the work necessary to
4999 deduce information about a private key may be performed offline. The amount
5000 of resources required for such an attack would be very significant and
5001 likely only accessible to a limited number of attackers. An attacker would
5002 additionally need online access to an unpatched system using the target
5003 private key in a scenario with persistent DH parameters and a private
5004 key that is shared between multiple clients. For example this can occur by
5005 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5006 similar to CVE-2015-3193 but must be treated as a separate problem.
5008 This issue was reported to OpenSSL by the OSS-Fuzz project.
5013 * Montgomery multiplication may produce incorrect results
5015 There is a carry propagating bug in the Broadwell-specific Montgomery
5016 multiplication procedure that handles input lengths divisible by, but
5017 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5018 and DH private keys are impossible. This is because the subroutine in
5019 question is not used in operations with the private key itself and an input
5020 of the attacker's direct choice. Otherwise the bug can manifest itself as
5021 transient authentication and key negotiation failures or reproducible
5022 erroneous outcome of public-key operations with specially crafted input.
5023 Among EC algorithms only Brainpool P-512 curves are affected and one
5024 presumably can attack ECDH key negotiation. Impact was not analyzed in
5025 detail, because pre-requisites for attack are considered unlikely. Namely
5026 multiple clients have to choose the curve in question and the server has to
5027 share the private key among them, neither of which is default behaviour.
5028 Even then only clients that chose the curve will be affected.
5030 This issue was publicly reported as transient failures and was not
5031 initially recognized as a security issue. Thanks to Richard Morgan for
5032 providing reproducible case.
5037 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5038 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5039 prevent issues where no progress is being made and the peer continually
5040 sends unrecognised record types, using up resources processing them.
5044 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5046 * Missing CRL sanity check
5048 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5049 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5050 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5052 This issue only affects the OpenSSL 1.0.2i
5057 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5059 * OCSP Status Request extension unbounded memory growth
5061 A malicious client can send an excessively large OCSP Status Request
5062 extension. If that client continually requests renegotiation, sending a
5063 large OCSP Status Request extension each time, then there will be unbounded
5064 memory growth on the server. This will eventually lead to a Denial Of
5065 Service attack through memory exhaustion. Servers with a default
5066 configuration are vulnerable even if they do not support OCSP. Builds using
5067 the "no-ocsp" build time option are not affected.
5069 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5074 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5077 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5083 * OOB write in MDC2_Update()
5085 An overflow can occur in MDC2_Update() either if called directly or
5086 through the EVP_DigestUpdate() function using MDC2. If an attacker
5087 is able to supply very large amounts of input data after a previous
5088 call to EVP_EncryptUpdate() with a partial block then a length check
5089 can overflow resulting in a heap corruption.
5091 The amount of data needed is comparable to SIZE_MAX which is impractical
5094 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5099 * Malformed SHA512 ticket DoS
5101 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5102 DoS attack where a malformed ticket will result in an OOB read which will
5105 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5106 a custom server callback and ticket lookup mechanism.
5108 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5113 * OOB write in BN_bn2dec()
5115 The function BN_bn2dec() does not check the return value of BN_div_word().
5116 This can cause an OOB write if an application uses this function with an
5117 overly large BIGNUM. This could be a problem if an overly large certificate
5118 or CRL is printed out from an untrusted source. TLS is not affected because
5119 record limits will reject an oversized certificate before it is parsed.
5121 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5126 * OOB read in TS_OBJ_print_bio()
5128 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5129 the total length the OID text representation would use and not the amount
5130 of data written. This will result in OOB reads when large OIDs are
5133 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5138 * Pointer arithmetic undefined behaviour
5140 Avoid some undefined pointer arithmetic
5142 A common idiom in the codebase is to check limits in the following manner:
5145 Where "p" points to some malloc'd data of SIZE bytes and
5148 "len" here could be from some externally supplied data (e.g. from a TLS
5151 The rules of C pointer arithmetic are such that "p + len" is only well
5152 defined where len <= SIZE. Therefore the above idiom is actually
5153 undefined behaviour.
5155 For example this could cause problems if some malloc implementation
5156 provides an address for "p" such that "p + len" actually overflows for
5157 values of len that are too big and therefore p + len < limit.
5159 This issue was reported to OpenSSL by Guido Vranken
5164 * Constant time flag not preserved in DSA signing
5166 Operations in the DSA signing algorithm should run in constant time in
5167 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5168 implementation means that a non-constant time codepath is followed for
5169 certain operations. This has been demonstrated through a cache-timing
5170 attack to be sufficient for an attacker to recover the private DSA key.
5172 This issue was reported by César Pereida (Aalto University), Billy Brumley
5173 (Tampere University of Technology), and Yuval Yarom (The University of
5174 Adelaide and NICTA).
5179 * DTLS buffered message DoS
5181 In a DTLS connection where handshake messages are delivered out-of-order
5182 those messages that OpenSSL is not yet ready to process will be buffered
5183 for later use. Under certain circumstances, a flaw in the logic means that
5184 those messages do not get removed from the buffer even though the handshake
5185 has been completed. An attacker could force up to approx. 15 messages to
5186 remain in the buffer when they are no longer required. These messages will
5187 be cleared when the DTLS connection is closed. The default maximum size for
5188 a message is 100k. Therefore the attacker could force an additional 1500k
5189 to be consumed per connection. By opening many simulataneous connections an
5190 attacker could cause a DoS attack through memory exhaustion.
5192 This issue was reported to OpenSSL by Quan Luo.
5197 * DTLS replay protection DoS
5199 A flaw in the DTLS replay attack protection mechanism means that records
5200 that arrive for future epochs update the replay protection "window" before
5201 the MAC for the record has been validated. This could be exploited by an
5202 attacker by sending a record for the next epoch (which does not have to
5203 decrypt or have a valid MAC), with a very large sequence number. This means
5204 that all subsequent legitimate packets are dropped causing a denial of
5205 service for a specific DTLS connection.
5207 This issue was reported to OpenSSL by the OCAP audit team.
5212 * Certificate message OOB reads
5214 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5215 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5216 theoretical DoS risk but this has not been observed in practice on common
5219 The messages affected are client certificate, client certificate request
5220 and server certificate. As a result the attack can only be performed
5221 against a client or a server which enables client authentication.
5223 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5228 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5230 * Prevent padding oracle in AES-NI CBC MAC check
5232 A MITM attacker can use a padding oracle attack to decrypt traffic
5233 when the connection uses an AES CBC cipher and the server support
5236 This issue was introduced as part of the fix for Lucky 13 padding
5237 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5238 constant time by making sure that always the same bytes are read and
5239 compared against either the MAC or padding bytes. But it no longer
5240 checked that there was enough data to have both the MAC and padding
5243 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5247 * Fix EVP_EncodeUpdate overflow
5249 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5250 Base64 encoding of binary data. If an attacker is able to supply very large
5251 amounts of input data then a length check can overflow resulting in a heap
5254 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5255 the `PEM_write_bio*` family of functions. These are mainly used within the
5256 OpenSSL command line applications, so any application which processes data
5257 from an untrusted source and outputs it as a PEM file should be considered
5258 vulnerable to this issue. User applications that call these APIs directly
5259 with large amounts of untrusted data may also be vulnerable.
5261 This issue was reported by Guido Vranken.
5266 * Fix EVP_EncryptUpdate overflow
5268 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5269 is able to supply very large amounts of input data after a previous call to
5270 EVP_EncryptUpdate() with a partial block then a length check can overflow
5271 resulting in a heap corruption. Following an analysis of all OpenSSL
5272 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5273 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5274 the first called function after an EVP_EncryptInit(), and therefore that
5275 specific call must be safe. The second form is where the length passed to
5276 EVP_EncryptUpdate() can be seen from the code to be some small value and
5277 therefore there is no possibility of an overflow. Since all instances are
5278 one of these two forms, it is believed that there can be no overflows in
5279 internal code due to this problem. It should be noted that
5280 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5281 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5282 of these calls have also been analysed too and it is believed there are no
5283 instances in internal usage where an overflow could occur.
5285 This issue was reported by Guido Vranken.
5290 * Prevent ASN.1 BIO excessive memory allocation
5292 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5293 a short invalid encoding can cause allocation of large amounts of memory
5294 potentially consuming excessive resources or exhausting memory.
5296 Any application parsing untrusted data through d2i BIO functions is
5297 affected. The memory based functions such as d2i_X509() are *not* affected.
5298 Since the memory based functions are used by the TLS library, TLS
5299 applications are not affected.
5301 This issue was reported by Brian Carpenter.
5308 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5309 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5310 in arbitrary stack data being returned in the buffer.
5312 This issue was reported by Guido Vranken.
5317 * Modify behavior of ALPN to invoke callback after SNI/servername
5318 callback, such that updates to the SSL_CTX affect ALPN.
5322 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5327 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5328 methods are enabled and ssl2 is disabled the methods return NULL.
5332 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5334 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5335 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5336 provide any "EXPORT" or "LOW" strength ciphers.
5340 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5341 is by default disabled at build-time. Builds that are not configured with
5342 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5343 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5344 will need to explicitly call either of:
5346 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5348 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5350 as appropriate. Even if either of those is used, or the application
5351 explicitly uses the version-specific SSLv2_method() or its client and
5352 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5353 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5354 ciphers, and SSLv2 56-bit DES are no longer available.
5359 * Fix a double-free in DSA code
5361 A double free bug was discovered when OpenSSL parses malformed DSA private
5362 keys and could lead to a DoS attack or memory corruption for applications
5363 that receive DSA private keys from untrusted sources. This scenario is
5366 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5372 * Disable SRP fake user seed to address a server memory leak.
5374 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5376 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5377 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5378 was changed to ignore the "fake user" SRP seed, even if the seed
5381 Users should use SRP_VBASE_get1_by_user instead. Note that in
5382 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5383 also that even though configuring the SRP seed attempts to hide
5384 invalid usernames by continuing the handshake with fake
5385 credentials, this behaviour is not constant time and no strong
5386 guarantees are made that the handshake is indistinguishable from
5387 that of a valid user.
5392 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5394 In the BN_hex2bn function the number of hex digits is calculated using an
5395 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5396 large values of `i` this can result in `bn_expand` not allocating any
5397 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5398 field as NULL leading to a subsequent NULL ptr deref. For very large values
5399 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5400 In this case memory is allocated to the internal BIGNUM data field, but it
5401 is insufficiently sized leading to heap corruption. A similar issue exists
5402 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5403 is ever called by user applications with very large untrusted hex/dec data.
5404 This is anticipated to be a rare occurrence.
5406 All OpenSSL internal usage of these functions use data that is not expected
5407 to be untrusted, e.g. config file data or application command line
5408 arguments. If user developed applications generate config file data based
5409 on untrusted data then it is possible that this could also lead to security
5410 consequences. This is also anticipated to be rare.
5412 This issue was reported to OpenSSL by Guido Vranken.
5417 * Fix memory issues in `BIO_*printf` functions
5419 The internal `fmtstr` function used in processing a "%s" format string in
5420 the `BIO_*printf` functions could overflow while calculating the length of a
5421 string and cause an OOB read when printing very long strings.
5423 Additionally the internal `doapr_outch` function can attempt to write to an
5424 OOB memory location (at an offset from the NULL pointer) in the event of a
5425 memory allocation failure. In 1.0.2 and below this could be caused where
5426 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5427 could be in processing a very long "%s" format string. Memory leaks can
5430 The first issue may mask the second issue dependent on compiler behaviour.
5431 These problems could enable attacks where large amounts of untrusted data
5432 is passed to the `BIO_*printf` functions. If applications use these functions
5433 in this way then they could be vulnerable. OpenSSL itself uses these
5434 functions when printing out human-readable dumps of ASN.1 data. Therefore
5435 applications that print this data could be vulnerable if the data is from
5436 untrusted sources. OpenSSL command line applications could also be
5437 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5438 as command line arguments.
5440 Libssl is not considered directly vulnerable. Additionally certificates etc
5441 received via remote connections via libssl are also unlikely to be able to
5442 trigger these issues because of message size limits enforced within libssl.
5444 This issue was reported to OpenSSL Guido Vranken.
5449 * Side channel attack on modular exponentiation
5451 A side-channel attack was found which makes use of cache-bank conflicts on
5452 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5453 of RSA keys. The ability to exploit this issue is limited as it relies on
5454 an attacker who has control of code in a thread running on the same
5455 hyper-threaded core as the victim thread which is performing decryptions.
5457 This issue was reported to OpenSSL by Yuval Yarom, The University of
5458 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5459 Nadia Heninger, University of Pennsylvania with more information at
5460 <http://cachebleed.info>.
5465 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5466 if no keysize is specified with default_bits. This fixes an
5467 omission in an earlier change that changed all RSA/DSA key generation
5468 commands to use 2048 bits by default.
5472 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5474 * DH small subgroups
5476 Historically OpenSSL only ever generated DH parameters based on "safe"
5477 primes. More recently (in version 1.0.2) support was provided for
5478 generating X9.42 style parameter files such as those required for RFC 5114
5479 support. The primes used in such files may not be "safe". Where an
5480 application is using DH configured with parameters based on primes that are
5481 not "safe" then an attacker could use this fact to find a peer's private
5482 DH exponent. This attack requires that the attacker complete multiple
5483 handshakes in which the peer uses the same private DH exponent. For example
5484 this could be used to discover a TLS server's private DH exponent if it's
5485 reusing the private DH exponent or it's using a static DH ciphersuite.
5487 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5488 TLS. It is not on by default. If the option is not set then the server
5489 reuses the same private DH exponent for the life of the server process and
5490 would be vulnerable to this attack. It is believed that many popular
5491 applications do set this option and would therefore not be at risk.
5493 The fix for this issue adds an additional check where a "q" parameter is
5494 available (as is the case in X9.42 based parameters). This detects the
5495 only known attack, and is the only possible defense for static DH
5496 ciphersuites. This could have some performance impact.
5498 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5499 default and cannot be disabled. This could have some performance impact.
5501 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5506 * SSLv2 doesn't block disabled ciphers
5508 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5509 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5510 been disabled, provided that the SSLv2 protocol was not also disabled via
5513 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5514 and Sebastian Schinzel.
5519 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5521 * BN_mod_exp may produce incorrect results on x86_64
5523 There is a carry propagating bug in the x86_64 Montgomery squaring
5524 procedure. No EC algorithms are affected. Analysis suggests that attacks
5525 against RSA and DSA as a result of this defect would be very difficult to
5526 perform and are not believed likely. Attacks against DH are considered just
5527 feasible (although very difficult) because most of the work necessary to
5528 deduce information about a private key may be performed offline. The amount
5529 of resources required for such an attack would be very significant and
5530 likely only accessible to a limited number of attackers. An attacker would
5531 additionally need online access to an unpatched system using the target
5532 private key in a scenario with persistent DH parameters and a private
5533 key that is shared between multiple clients. For example this can occur by
5534 default in OpenSSL DHE based SSL/TLS ciphersuites.
5536 This issue was reported to OpenSSL by Hanno Böck.
5541 * Certificate verify crash with missing PSS parameter
5543 The signature verification routines will crash with a NULL pointer
5544 dereference if presented with an ASN.1 signature using the RSA PSS
5545 algorithm and absent mask generation function parameter. Since these
5546 routines are used to verify certificate signature algorithms this can be
5547 used to crash any certificate verification operation and exploited in a
5548 DoS attack. Any application which performs certificate verification is
5549 vulnerable including OpenSSL clients and servers which enable client
5552 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5557 * X509_ATTRIBUTE memory leak
5559 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5560 memory. This structure is used by the PKCS#7 and CMS routines so any
5561 application which reads PKCS#7 or CMS data from untrusted sources is
5562 affected. SSL/TLS is not affected.
5564 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5570 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5571 This changes the decoding behaviour for some invalid messages,
5572 though the change is mostly in the more lenient direction, and
5573 legacy behaviour is preserved as much as possible.
5577 * In DSA_generate_parameters_ex, if the provided seed is too short,
5580 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5582 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5584 * Alternate chains certificate forgery
5586 During certificate verification, OpenSSL will attempt to find an
5587 alternative certificate chain if the first attempt to build such a chain
5588 fails. An error in the implementation of this logic can mean that an
5589 attacker could cause certain checks on untrusted certificates to be
5590 bypassed, such as the CA flag, enabling them to use a valid leaf
5591 certificate to act as a CA and "issue" an invalid certificate.
5593 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5598 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5600 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5601 incompatibility in the handling of HMAC. The previous ABI has now been
5606 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5608 * Malformed ECParameters causes infinite loop
5610 When processing an ECParameters structure OpenSSL enters an infinite loop
5611 if the curve specified is over a specially malformed binary polynomial
5614 This can be used to perform denial of service against any
5615 system which processes public keys, certificate requests or
5616 certificates. This includes TLS clients and TLS servers with
5617 client authentication enabled.
5619 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5624 * Exploitable out-of-bounds read in X509_cmp_time
5626 X509_cmp_time does not properly check the length of the ASN1_TIME
5627 string and can read a few bytes out of bounds. In addition,
5628 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5631 An attacker can use this to craft malformed certificates and CRLs of
5632 various sizes and potentially cause a segmentation fault, resulting in
5633 a DoS on applications that verify certificates or CRLs. TLS clients
5634 that verify CRLs are affected. TLS clients and servers with client
5635 authentication enabled may be affected if they use custom verification
5638 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5639 independently by Hanno Böck.
5644 * PKCS7 crash with missing EnvelopedContent
5646 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5647 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5648 with missing content and trigger a NULL pointer dereference on parsing.
5650 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5651 structures from untrusted sources are affected. OpenSSL clients and
5652 servers are not affected.
5654 This issue was reported to OpenSSL by Michal Zalewski (Google).
5659 * CMS verify infinite loop with unknown hash function
5661 When verifying a signedData message the CMS code can enter an infinite loop
5662 if presented with an unknown hash function OID. This can be used to perform
5663 denial of service against any system which verifies signedData messages using
5665 This issue was reported to OpenSSL by Johannes Bauer.
5670 * Race condition handling NewSessionTicket
5672 If a NewSessionTicket is received by a multi-threaded client when attempting to
5673 reuse a previous ticket then a race condition can occur potentially leading to
5674 a double free of the ticket data.
5679 * Only support 256-bit or stronger elliptic curves with the
5680 'ecdh_auto' setting (server) or by default (client). Of supported
5681 curves, prefer P-256 (both).
5685 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5687 * ClientHello sigalgs DoS fix
5689 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5690 invalid signature algorithms extension a NULL pointer dereference will
5691 occur. This can be exploited in a DoS attack against the server.
5693 This issue was was reported to OpenSSL by David Ramos of Stanford
5697 *Stephen Henson and Matt Caswell*
5699 * Multiblock corrupted pointer fix
5701 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5702 feature only applies on 64 bit x86 architecture platforms that support AES
5703 NI instructions. A defect in the implementation of "multiblock" can cause
5704 OpenSSL's internal write buffer to become incorrectly set to NULL when
5705 using non-blocking IO. Typically, when the user application is using a
5706 socket BIO for writing, this will only result in a failed connection.
5707 However if some other BIO is used then it is likely that a segmentation
5708 fault will be triggered, thus enabling a potential DoS attack.
5710 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5715 * Segmentation fault in DTLSv1_listen fix
5717 The DTLSv1_listen function is intended to be stateless and processes the
5718 initial ClientHello from many peers. It is common for user code to loop
5719 over the call to DTLSv1_listen until a valid ClientHello is received with
5720 an associated cookie. A defect in the implementation of DTLSv1_listen means
5721 that state is preserved in the SSL object from one invocation to the next
5722 that can lead to a segmentation fault. Errors processing the initial
5723 ClientHello can trigger this scenario. An example of such an error could be
5724 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5727 This issue was reported to OpenSSL by Per Allansson.
5732 * Segmentation fault in ASN1_TYPE_cmp fix
5734 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5735 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5736 certificate signature algorithm consistency this can be used to crash any
5737 certificate verification operation and exploited in a DoS attack. Any
5738 application which performs certificate verification is vulnerable including
5739 OpenSSL clients and servers which enable client authentication.
5744 * Segmentation fault for invalid PSS parameters fix
5746 The signature verification routines will crash with a NULL pointer
5747 dereference if presented with an ASN.1 signature using the RSA PSS
5748 algorithm and invalid parameters. Since these routines are used to verify
5749 certificate signature algorithms this can be used to crash any
5750 certificate verification operation and exploited in a DoS attack. Any
5751 application which performs certificate verification is vulnerable including
5752 OpenSSL clients and servers which enable client authentication.
5754 This issue was was reported to OpenSSL by Brian Carpenter.
5759 * ASN.1 structure reuse memory corruption fix
5761 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5762 memory corruption via an invalid write. Such reuse is and has been
5763 strongly discouraged and is believed to be rare.
5765 Applications that parse structures containing CHOICE or ANY DEFINED BY
5766 components may be affected. Certificate parsing (d2i_X509 and related
5767 functions) are however not affected. OpenSSL clients and servers are
5773 * PKCS7 NULL pointer dereferences fix
5775 The PKCS#7 parsing code does not handle missing outer ContentInfo
5776 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5777 missing content and trigger a NULL pointer dereference on parsing.
5779 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5780 otherwise parse PKCS#7 structures from untrusted sources are
5781 affected. OpenSSL clients and servers are not affected.
5783 This issue was reported to OpenSSL by Michal Zalewski (Google).
5788 * DoS via reachable assert in SSLv2 servers fix
5790 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5791 servers that both support SSLv2 and enable export cipher suites by sending
5792 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5794 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5795 (OpenSSL development team).
5800 * Empty CKE with client auth and DHE fix
5802 If client auth is used then a server can seg fault in the event of a DHE
5803 ciphersuite being selected and a zero length ClientKeyExchange message
5804 being sent by the client. This could be exploited in a DoS attack.
5809 * Handshake with unseeded PRNG fix
5811 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5812 with an unseeded PRNG. The conditions are:
5813 - The client is on a platform where the PRNG has not been seeded
5814 automatically, and the user has not seeded manually
5815 - A protocol specific client method version has been used (i.e. not
5816 SSL_client_methodv23)
5817 - A ciphersuite is used that does not require additional random data from
5818 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5820 If the handshake succeeds then the client random that has been used will
5821 have been generated from a PRNG with insufficient entropy and therefore the
5822 output may be predictable.
5824 For example using the following command with an unseeded openssl will
5825 succeed on an unpatched platform:
5827 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5832 * Use After Free following d2i_ECPrivatekey error fix
5834 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5835 could cause a use after free condition. This, in turn, could cause a double
5836 free in several private key parsing functions (such as d2i_PrivateKey
5837 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5838 for applications that receive EC private keys from untrusted
5839 sources. This scenario is considered rare.
5841 This issue was discovered by the BoringSSL project and fixed in their
5847 * X509_to_X509_REQ NULL pointer deref fix
5849 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5850 the certificate key is invalid. This function is rarely used in practice.
5852 This issue was discovered by Brian Carpenter.
5857 * Removed the export ciphers from the DEFAULT ciphers
5861 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5863 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5864 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5865 So far those who have to target multiple platforms would compromise
5866 and argue that binary targeting say ARMv5 would still execute on
5867 ARMv8. "Universal" build resolves this compromise by providing
5868 near-optimal performance even on newer platforms.
5872 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5873 (other platforms pending).
5875 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5877 * Add support for the SignedCertificateTimestampList certificate and
5878 OCSP response extensions from RFC6962.
5882 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5883 for corner cases. (Certain input points at infinity could lead to
5884 bogus results, with non-infinity inputs mapped to infinity too.)
5888 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5889 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5890 common cases are optimized and there still is room for further
5891 improvements. Vector Permutation AES for Altivec is also added.
5895 * Add support for little-endian ppc64 Linux target.
5897 *Marcelo Cerri (IBM)*
5899 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5900 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5901 are optimized and there still is room for further improvements.
5902 Both 32- and 64-bit modes are supported.
5904 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5906 * Improved ARMv7 NEON support.
5910 * Support for SPARC Architecture 2011 crypto extensions, first
5911 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5912 SHA256/512, MD5, GHASH and modular exponentiation.
5914 *Andy Polyakov, David Miller*
5916 * Accelerated modular exponentiation for Intel processors, a.k.a.
5919 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5921 * Support for new and upcoming Intel processors, including AVX2,
5922 BMI and SHA ISA extensions. This includes additional "stitched"
5923 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5926 This work was sponsored by Intel Corp.
5930 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5931 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5932 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5936 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5937 this fixes a limitation in previous versions of OpenSSL.
5941 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5942 MGF1 digest and OAEP label.
5946 * Add EVP support for key wrapping algorithms, to avoid problems with
5947 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5948 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5949 algorithms and include tests cases.
5953 * Add functions to allocate and set the fields of an ECDSA_METHOD
5956 *Douglas E. Engert, Steve Henson*
5958 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5959 difference in days and seconds between two tm or ASN1_TIME structures.
5963 * Add -rev test option to s_server to just reverse order of characters
5964 received by client and send back to server. Also prints an abbreviated
5965 summary of the connection parameters.
5969 * New option -brief for s_client and s_server to print out a brief summary
5970 of connection parameters.
5974 * Add callbacks for arbitrary TLS extensions.
5976 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5978 * New option -crl_download in several openssl utilities to download CRLs
5979 from CRLDP extension in certificates.
5983 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5987 * New function X509_CRL_diff to generate a delta CRL from the difference
5988 of two full CRLs. Add support to "crl" utility.
5992 * New functions to set lookup_crls function and to retrieve
5993 X509_STORE from X509_STORE_CTX.
5997 * Print out deprecated issuer and subject unique ID fields in
6002 * Extend OCSP I/O functions so they can be used for simple general purpose
6003 HTTP as well as OCSP. New wrapper function which can be used to download
6004 CRLs using the OCSP API.
6008 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6012 * `SSL_CONF*` functions. These provide a common framework for application
6013 configuration using configuration files or command lines.
6017 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6018 message callback and prints the results. Needs compile time option
6019 "enable-ssl-trace". New options to s_client and s_server to enable
6024 * New ctrl and macro to retrieve supported points extensions.
6025 Print out extension in s_server and s_client.
6029 * New functions to retrieve certificate signature and signature
6034 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6039 * New Suite B modes for TLS code. These use and enforce the requirements
6040 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6041 only use Suite B curves. The Suite B modes can be set by using the
6042 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6046 * New chain verification flags for Suite B levels of security. Check
6047 algorithms are acceptable when flags are set in X509_verify_cert.
6051 * Make tls1_check_chain return a set of flags indicating checks passed
6052 by a certificate chain. Add additional tests to handle client
6053 certificates: checks for matching certificate type and issuer name
6058 * If an attempt is made to use a signature algorithm not in the peer
6059 preference list abort the handshake. If client has no suitable
6060 signature algorithms in response to a certificate request do not
6061 use the certificate.
6065 * If server EC tmp key is not in client preference list abort handshake.
6069 * Add support for certificate stores in CERT structure. This makes it
6070 possible to have different stores per SSL structure or one store in
6071 the parent SSL_CTX. Include distinct stores for certificate chain
6072 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6073 to build and store a certificate chain in CERT structure: returning
6074 an error if the chain cannot be built: this will allow applications
6075 to test if a chain is correctly configured.
6077 Note: if the CERT based stores are not set then the parent SSL_CTX
6078 store is used to retain compatibility with existing behaviour.
6082 * New function ssl_set_client_disabled to set a ciphersuite disabled
6083 mask based on the current session, check mask when sending client
6084 hello and checking the requested ciphersuite.
6088 * New ctrls to retrieve and set certificate types in a certificate
6089 request message. Print out received values in s_client. If certificate
6090 types is not set with custom values set sensible values based on
6091 supported signature algorithms.
6095 * Support for distinct client and server supported signature algorithms.
6099 * Add certificate callback. If set this is called whenever a certificate
6100 is required by client or server. An application can decide which
6101 certificate chain to present based on arbitrary criteria: for example
6102 supported signature algorithms. Add very simple example to s_server.
6103 This fixes many of the problems and restrictions of the existing client
6104 certificate callback: for example you can now clear an existing
6105 certificate and specify the whole chain.
6109 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6110 the certificate can be used for (if anything). Set valid_flags field
6111 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6112 to have similar checks in it.
6114 Add new "cert_flags" field to CERT structure and include a "strict mode".
6115 This enforces some TLS certificate requirements (such as only permitting
6116 certificate signature algorithms contained in the supported algorithms
6117 extension) which some implementations ignore: this option should be used
6118 with caution as it could cause interoperability issues.
6122 * Update and tidy signature algorithm extension processing. Work out
6123 shared signature algorithms based on preferences and peer algorithms
6124 and print them out in s_client and s_server. Abort handshake if no
6125 shared signature algorithms.
6129 * Add new functions to allow customised supported signature algorithms
6130 for SSL and SSL_CTX structures. Add options to s_client and s_server
6135 * New function SSL_certs_clear() to delete all references to certificates
6136 from an SSL structure. Before this once a certificate had been added
6137 it couldn't be removed.
6141 * Integrate hostname, email address and IP address checking with certificate
6142 verification. New verify options supporting checking in openssl utility.
6146 * Fixes and wildcard matching support to hostname and email checking
6147 functions. Add manual page.
6149 *Florian Weimer (Red Hat Product Security Team)*
6151 * New functions to check a hostname email or IP address against a
6152 certificate. Add options x509 utility to print results of checks against
6157 * Fix OCSP checking.
6159 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6161 * Initial experimental support for explicitly trusted non-root CAs.
6162 OpenSSL still tries to build a complete chain to a root but if an
6163 intermediate CA has a trust setting included that is used. The first
6164 setting is used: whether to trust (e.g., -addtrust option to the x509
6169 * Add -trusted_first option which attempts to find certificates in the
6170 trusted store even if an untrusted chain is also supplied.
6174 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6175 platform support for Linux and Android.
6179 * Support for linux-x32, ILP32 environment in x86_64 framework.
6183 * Experimental multi-implementation support for FIPS capable OpenSSL.
6184 When in FIPS mode the approved implementations are used as normal,
6185 when not in FIPS mode the internal unapproved versions are used instead.
6186 This means that the FIPS capable OpenSSL isn't forced to use the
6187 (often lower performance) FIPS implementations outside FIPS mode.
6191 * Transparently support X9.42 DH parameters when calling
6192 PEM_read_bio_DHparameters. This means existing applications can handle
6193 the new parameter format automatically.
6197 * Initial experimental support for X9.42 DH parameter format: mainly
6198 to support use of 'q' parameter for RFC5114 parameters.
6202 * Add DH parameters from RFC5114 including test data to dhtest.
6206 * Support for automatic EC temporary key parameter selection. If enabled
6207 the most preferred EC parameters are automatically used instead of
6208 hardcoded fixed parameters. Now a server just has to call:
6209 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6210 support ECDH and use the most appropriate parameters.
6214 * Enhance and tidy EC curve and point format TLS extension code. Use
6215 static structures instead of allocation if default values are used.
6216 New ctrls to set curves we wish to support and to retrieve shared curves.
6217 Print out shared curves in s_server. New options to s_server and s_client
6218 to set list of supported curves.
6222 * New ctrls to retrieve supported signature algorithms and
6223 supported curve values as an array of NIDs. Extend openssl utility
6224 to print out received values.
6228 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6229 between NIDs and the more common NIST names such as "P-256". Enhance
6230 ecparam utility and ECC method to recognise the NIST names for curves.
6234 * Enhance SSL/TLS certificate chain handling to support different
6235 chains for each certificate instead of one chain in the parent SSL_CTX.
6239 * Support for fixed DH ciphersuite client authentication: where both
6240 server and client use DH certificates with common parameters.
6244 * Support for fixed DH ciphersuites: those requiring DH server
6249 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6251 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6252 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6253 X509_CINF_get_signature were reverted post internal team review.
6258 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6260 * OCSP Status Request extension unbounded memory growth
6262 A malicious client can send an excessively large OCSP Status Request
6263 extension. If that client continually requests renegotiation, sending a
6264 large OCSP Status Request extension each time, then there will be unbounded
6265 memory growth on the server. This will eventually lead to a Denial Of
6266 Service attack through memory exhaustion. Servers with a default
6267 configuration are vulnerable even if they do not support OCSP. Builds using
6268 the "no-ocsp" build time option are not affected.
6270 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6275 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6278 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6284 * OOB write in MDC2_Update()
6286 An overflow can occur in MDC2_Update() either if called directly or
6287 through the EVP_DigestUpdate() function using MDC2. If an attacker
6288 is able to supply very large amounts of input data after a previous
6289 call to EVP_EncryptUpdate() with a partial block then a length check
6290 can overflow resulting in a heap corruption.
6292 The amount of data needed is comparable to SIZE_MAX which is impractical
6295 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6300 * Malformed SHA512 ticket DoS
6302 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6303 DoS attack where a malformed ticket will result in an OOB read which will
6306 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6307 a custom server callback and ticket lookup mechanism.
6309 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6314 * OOB write in BN_bn2dec()
6316 The function BN_bn2dec() does not check the return value of BN_div_word().
6317 This can cause an OOB write if an application uses this function with an
6318 overly large BIGNUM. This could be a problem if an overly large certificate
6319 or CRL is printed out from an untrusted source. TLS is not affected because
6320 record limits will reject an oversized certificate before it is parsed.
6322 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6327 * OOB read in TS_OBJ_print_bio()
6329 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6330 the total length the OID text representation would use and not the amount
6331 of data written. This will result in OOB reads when large OIDs are
6334 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6339 * Pointer arithmetic undefined behaviour
6341 Avoid some undefined pointer arithmetic
6343 A common idiom in the codebase is to check limits in the following manner:
6346 Where "p" points to some malloc'd data of SIZE bytes and
6349 "len" here could be from some externally supplied data (e.g. from a TLS
6352 The rules of C pointer arithmetic are such that "p + len" is only well
6353 defined where len <= SIZE. Therefore the above idiom is actually
6354 undefined behaviour.
6356 For example this could cause problems if some malloc implementation
6357 provides an address for "p" such that "p + len" actually overflows for
6358 values of len that are too big and therefore p + len < limit.
6360 This issue was reported to OpenSSL by Guido Vranken
6365 * Constant time flag not preserved in DSA signing
6367 Operations in the DSA signing algorithm should run in constant time in
6368 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6369 implementation means that a non-constant time codepath is followed for
6370 certain operations. This has been demonstrated through a cache-timing
6371 attack to be sufficient for an attacker to recover the private DSA key.
6373 This issue was reported by César Pereida (Aalto University), Billy Brumley
6374 (Tampere University of Technology), and Yuval Yarom (The University of
6375 Adelaide and NICTA).
6380 * DTLS buffered message DoS
6382 In a DTLS connection where handshake messages are delivered out-of-order
6383 those messages that OpenSSL is not yet ready to process will be buffered
6384 for later use. Under certain circumstances, a flaw in the logic means that
6385 those messages do not get removed from the buffer even though the handshake
6386 has been completed. An attacker could force up to approx. 15 messages to
6387 remain in the buffer when they are no longer required. These messages will
6388 be cleared when the DTLS connection is closed. The default maximum size for
6389 a message is 100k. Therefore the attacker could force an additional 1500k
6390 to be consumed per connection. By opening many simulataneous connections an
6391 attacker could cause a DoS attack through memory exhaustion.
6393 This issue was reported to OpenSSL by Quan Luo.
6398 * DTLS replay protection DoS
6400 A flaw in the DTLS replay attack protection mechanism means that records
6401 that arrive for future epochs update the replay protection "window" before
6402 the MAC for the record has been validated. This could be exploited by an
6403 attacker by sending a record for the next epoch (which does not have to
6404 decrypt or have a valid MAC), with a very large sequence number. This means
6405 that all subsequent legitimate packets are dropped causing a denial of
6406 service for a specific DTLS connection.
6408 This issue was reported to OpenSSL by the OCAP audit team.
6413 * Certificate message OOB reads
6415 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6416 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6417 theoretical DoS risk but this has not been observed in practice on common
6420 The messages affected are client certificate, client certificate request
6421 and server certificate. As a result the attack can only be performed
6422 against a client or a server which enables client authentication.
6424 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6429 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6431 * Prevent padding oracle in AES-NI CBC MAC check
6433 A MITM attacker can use a padding oracle attack to decrypt traffic
6434 when the connection uses an AES CBC cipher and the server support
6437 This issue was introduced as part of the fix for Lucky 13 padding
6438 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6439 constant time by making sure that always the same bytes are read and
6440 compared against either the MAC or padding bytes. But it no longer
6441 checked that there was enough data to have both the MAC and padding
6444 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6449 * Fix EVP_EncodeUpdate overflow
6451 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6452 Base64 encoding of binary data. If an attacker is able to supply very large
6453 amounts of input data then a length check can overflow resulting in a heap
6456 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6457 the `PEM_write_bio*` family of functions. These are mainly used within the
6458 OpenSSL command line applications, so any application which processes data
6459 from an untrusted source and outputs it as a PEM file should be considered
6460 vulnerable to this issue. User applications that call these APIs directly
6461 with large amounts of untrusted data may also be vulnerable.
6463 This issue was reported by Guido Vranken.
6468 * Fix EVP_EncryptUpdate overflow
6470 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6471 is able to supply very large amounts of input data after a previous call to
6472 EVP_EncryptUpdate() with a partial block then a length check can overflow
6473 resulting in a heap corruption. Following an analysis of all OpenSSL
6474 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6475 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6476 the first called function after an EVP_EncryptInit(), and therefore that
6477 specific call must be safe. The second form is where the length passed to
6478 EVP_EncryptUpdate() can be seen from the code to be some small value and
6479 therefore there is no possibility of an overflow. Since all instances are
6480 one of these two forms, it is believed that there can be no overflows in
6481 internal code due to this problem. It should be noted that
6482 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6483 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6484 of these calls have also been analysed too and it is believed there are no
6485 instances in internal usage where an overflow could occur.
6487 This issue was reported by Guido Vranken.
6492 * Prevent ASN.1 BIO excessive memory allocation
6494 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6495 a short invalid encoding can casuse allocation of large amounts of memory
6496 potentially consuming excessive resources or exhausting memory.
6498 Any application parsing untrusted data through d2i BIO functions is
6499 affected. The memory based functions such as d2i_X509() are *not* affected.
6500 Since the memory based functions are used by the TLS library, TLS
6501 applications are not affected.
6503 This issue was reported by Brian Carpenter.
6510 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6511 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6512 in arbitrary stack data being returned in the buffer.
6514 This issue was reported by Guido Vranken.
6519 * Modify behavior of ALPN to invoke callback after SNI/servername
6520 callback, such that updates to the SSL_CTX affect ALPN.
6524 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6529 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6530 methods are enabled and ssl2 is disabled the methods return NULL.
6534 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6536 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6537 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6538 provide any "EXPORT" or "LOW" strength ciphers.
6542 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6543 is by default disabled at build-time. Builds that are not configured with
6544 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6545 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6546 will need to explicitly call either of:
6548 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6550 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6552 as appropriate. Even if either of those is used, or the application
6553 explicitly uses the version-specific SSLv2_method() or its client and
6554 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6555 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6556 ciphers, and SSLv2 56-bit DES are no longer available.
6561 * Fix a double-free in DSA code
6563 A double free bug was discovered when OpenSSL parses malformed DSA private
6564 keys and could lead to a DoS attack or memory corruption for applications
6565 that receive DSA private keys from untrusted sources. This scenario is
6568 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6574 * Disable SRP fake user seed to address a server memory leak.
6576 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6578 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6579 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6580 was changed to ignore the "fake user" SRP seed, even if the seed
6583 Users should use SRP_VBASE_get1_by_user instead. Note that in
6584 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6585 also that even though configuring the SRP seed attempts to hide
6586 invalid usernames by continuing the handshake with fake
6587 credentials, this behaviour is not constant time and no strong
6588 guarantees are made that the handshake is indistinguishable from
6589 that of a valid user.
6594 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6596 In the BN_hex2bn function the number of hex digits is calculated using an
6597 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6598 large values of `i` this can result in `bn_expand` not allocating any
6599 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6600 field as NULL leading to a subsequent NULL ptr deref. For very large values
6601 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6602 In this case memory is allocated to the internal BIGNUM data field, but it
6603 is insufficiently sized leading to heap corruption. A similar issue exists
6604 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6605 is ever called by user applications with very large untrusted hex/dec data.
6606 This is anticipated to be a rare occurrence.
6608 All OpenSSL internal usage of these functions use data that is not expected
6609 to be untrusted, e.g. config file data or application command line
6610 arguments. If user developed applications generate config file data based
6611 on untrusted data then it is possible that this could also lead to security
6612 consequences. This is also anticipated to be rare.
6614 This issue was reported to OpenSSL by Guido Vranken.
6619 * Fix memory issues in `BIO_*printf` functions
6621 The internal `fmtstr` function used in processing a "%s" format string in
6622 the `BIO_*printf` functions could overflow while calculating the length of a
6623 string and cause an OOB read when printing very long strings.
6625 Additionally the internal `doapr_outch` function can attempt to write to an
6626 OOB memory location (at an offset from the NULL pointer) in the event of a
6627 memory allocation failure. In 1.0.2 and below this could be caused where
6628 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6629 could be in processing a very long "%s" format string. Memory leaks can
6632 The first issue may mask the second issue dependent on compiler behaviour.
6633 These problems could enable attacks where large amounts of untrusted data
6634 is passed to the `BIO_*printf` functions. If applications use these functions
6635 in this way then they could be vulnerable. OpenSSL itself uses these
6636 functions when printing out human-readable dumps of ASN.1 data. Therefore
6637 applications that print this data could be vulnerable if the data is from
6638 untrusted sources. OpenSSL command line applications could also be
6639 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6640 as command line arguments.
6642 Libssl is not considered directly vulnerable. Additionally certificates etc
6643 received via remote connections via libssl are also unlikely to be able to
6644 trigger these issues because of message size limits enforced within libssl.
6646 This issue was reported to OpenSSL Guido Vranken.
6651 * Side channel attack on modular exponentiation
6653 A side-channel attack was found which makes use of cache-bank conflicts on
6654 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6655 of RSA keys. The ability to exploit this issue is limited as it relies on
6656 an attacker who has control of code in a thread running on the same
6657 hyper-threaded core as the victim thread which is performing decryptions.
6659 This issue was reported to OpenSSL by Yuval Yarom, The University of
6660 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6661 Nadia Heninger, University of Pennsylvania with more information at
6662 <http://cachebleed.info>.
6667 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6668 if no keysize is specified with default_bits. This fixes an
6669 omission in an earlier change that changed all RSA/DSA key generation
6670 commands to use 2048 bits by default.
6674 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6676 * Protection for DH small subgroup attacks
6678 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6679 switched on by default and cannot be disabled. This could have some
6684 * SSLv2 doesn't block disabled ciphers
6686 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6687 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6688 been disabled, provided that the SSLv2 protocol was not also disabled via
6691 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6692 and Sebastian Schinzel.
6697 * Reject DH handshakes with parameters shorter than 1024 bits.
6701 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6703 * Certificate verify crash with missing PSS parameter
6705 The signature verification routines will crash with a NULL pointer
6706 dereference if presented with an ASN.1 signature using the RSA PSS
6707 algorithm and absent mask generation function parameter. Since these
6708 routines are used to verify certificate signature algorithms this can be
6709 used to crash any certificate verification operation and exploited in a
6710 DoS attack. Any application which performs certificate verification is
6711 vulnerable including OpenSSL clients and servers which enable client
6714 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6719 * X509_ATTRIBUTE memory leak
6721 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6722 memory. This structure is used by the PKCS#7 and CMS routines so any
6723 application which reads PKCS#7 or CMS data from untrusted sources is
6724 affected. SSL/TLS is not affected.
6726 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6732 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6733 This changes the decoding behaviour for some invalid messages,
6734 though the change is mostly in the more lenient direction, and
6735 legacy behaviour is preserved as much as possible.
6739 * In DSA_generate_parameters_ex, if the provided seed is too short,
6740 use a random seed, as already documented.
6742 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6744 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6746 * Alternate chains certificate forgery
6748 During certificate verfification, OpenSSL will attempt to find an
6749 alternative certificate chain if the first attempt to build such a chain
6750 fails. An error in the implementation of this logic can mean that an
6751 attacker could cause certain checks on untrusted certificates to be
6752 bypassed, such as the CA flag, enabling them to use a valid leaf
6753 certificate to act as a CA and "issue" an invalid certificate.
6755 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6761 * Race condition handling PSK identify hint
6763 If PSK identity hints are received by a multi-threaded client then
6764 the values are wrongly updated in the parent SSL_CTX structure. This can
6765 result in a race condition potentially leading to a double free of the
6771 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6773 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6774 incompatibility in the handling of HMAC. The previous ABI has now been
6777 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6779 * Malformed ECParameters causes infinite loop
6781 When processing an ECParameters structure OpenSSL enters an infinite loop
6782 if the curve specified is over a specially malformed binary polynomial
6785 This can be used to perform denial of service against any
6786 system which processes public keys, certificate requests or
6787 certificates. This includes TLS clients and TLS servers with
6788 client authentication enabled.
6790 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6795 * Exploitable out-of-bounds read in X509_cmp_time
6797 X509_cmp_time does not properly check the length of the ASN1_TIME
6798 string and can read a few bytes out of bounds. In addition,
6799 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6802 An attacker can use this to craft malformed certificates and CRLs of
6803 various sizes and potentially cause a segmentation fault, resulting in
6804 a DoS on applications that verify certificates or CRLs. TLS clients
6805 that verify CRLs are affected. TLS clients and servers with client
6806 authentication enabled may be affected if they use custom verification
6809 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6810 independently by Hanno Böck.
6815 * PKCS7 crash with missing EnvelopedContent
6817 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6818 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6819 with missing content and trigger a NULL pointer dereference on parsing.
6821 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6822 structures from untrusted sources are affected. OpenSSL clients and
6823 servers are not affected.
6825 This issue was reported to OpenSSL by Michal Zalewski (Google).
6830 * CMS verify infinite loop with unknown hash function
6832 When verifying a signedData message the CMS code can enter an infinite loop
6833 if presented with an unknown hash function OID. This can be used to perform
6834 denial of service against any system which verifies signedData messages using
6836 This issue was reported to OpenSSL by Johannes Bauer.
6841 * Race condition handling NewSessionTicket
6843 If a NewSessionTicket is received by a multi-threaded client when attempting to
6844 reuse a previous ticket then a race condition can occur potentially leading to
6845 a double free of the ticket data.
6850 * Reject DH handshakes with parameters shorter than 768 bits.
6852 *Kurt Roeckx and Emilia Kasper*
6854 * dhparam: generate 2048-bit parameters by default.
6856 *Kurt Roeckx and Emilia Kasper*
6858 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6860 * Segmentation fault in ASN1_TYPE_cmp fix
6862 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6863 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6864 certificate signature algorithm consistency this can be used to crash any
6865 certificate verification operation and exploited in a DoS attack. Any
6866 application which performs certificate verification is vulnerable including
6867 OpenSSL clients and servers which enable client authentication.
6872 * ASN.1 structure reuse memory corruption fix
6874 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6875 memory corruption via an invalid write. Such reuse is and has been
6876 strongly discouraged and is believed to be rare.
6878 Applications that parse structures containing CHOICE or ANY DEFINED BY
6879 components may be affected. Certificate parsing (d2i_X509 and related
6880 functions) are however not affected. OpenSSL clients and servers are
6886 * PKCS7 NULL pointer dereferences fix
6888 The PKCS#7 parsing code does not handle missing outer ContentInfo
6889 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6890 missing content and trigger a NULL pointer dereference on parsing.
6892 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6893 otherwise parse PKCS#7 structures from untrusted sources are
6894 affected. OpenSSL clients and servers are not affected.
6896 This issue was reported to OpenSSL by Michal Zalewski (Google).
6901 * DoS via reachable assert in SSLv2 servers fix
6903 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6904 servers that both support SSLv2 and enable export cipher suites by sending
6905 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6907 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6908 (OpenSSL development team).
6913 * Use After Free following d2i_ECPrivatekey error fix
6915 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6916 could cause a use after free condition. This, in turn, could cause a double
6917 free in several private key parsing functions (such as d2i_PrivateKey
6918 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6919 for applications that receive EC private keys from untrusted
6920 sources. This scenario is considered rare.
6922 This issue was discovered by the BoringSSL project and fixed in their
6928 * X509_to_X509_REQ NULL pointer deref fix
6930 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6931 the certificate key is invalid. This function is rarely used in practice.
6933 This issue was discovered by Brian Carpenter.
6938 * Removed the export ciphers from the DEFAULT ciphers
6942 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6944 * Build fixes for the Windows and OpenVMS platforms
6946 *Matt Caswell and Richard Levitte*
6948 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6950 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6951 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6952 dereference. This could lead to a Denial Of Service attack. Thanks to
6953 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6958 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6959 dtls1_buffer_record function under certain conditions. In particular this
6960 could occur if an attacker sent repeated DTLS records with the same
6961 sequence number but for the next epoch. The memory leak could be exploited
6962 by an attacker in a Denial of Service attack through memory exhaustion.
6963 Thanks to Chris Mueller for reporting this issue.
6968 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6969 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6970 method would be set to NULL which could later result in a NULL pointer
6971 dereference. Thanks to Frank Schmirler for reporting this issue.
6976 * Abort handshake if server key exchange message is omitted for ephemeral
6979 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6980 reporting this issue.
6985 * Remove non-export ephemeral RSA code on client and server. This code
6986 violated the TLS standard by allowing the use of temporary RSA keys in
6987 non-export ciphersuites and could be used by a server to effectively
6988 downgrade the RSA key length used to a value smaller than the server
6989 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6990 INRIA or reporting this issue.
6995 * Fixed issue where DH client certificates are accepted without verification.
6996 An OpenSSL server will accept a DH certificate for client authentication
6997 without the certificate verify message. This effectively allows a client to
6998 authenticate without the use of a private key. This only affects servers
6999 which trust a client certificate authority which issues certificates
7000 containing DH keys: these are extremely rare and hardly ever encountered.
7001 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7007 * Ensure that the session ID context of an SSL is updated when its
7008 SSL_CTX is updated via SSL_set_SSL_CTX.
7010 The session ID context is typically set from the parent SSL_CTX,
7011 and can vary with the CTX.
7015 * Fix various certificate fingerprint issues.
7017 By using non-DER or invalid encodings outside the signed portion of a
7018 certificate the fingerprint can be changed without breaking the signature.
7019 Although no details of the signed portion of the certificate can be changed
7020 this can cause problems with some applications: e.g. those using the
7021 certificate fingerprint for blacklists.
7023 1. Reject signatures with non zero unused bits.
7025 If the BIT STRING containing the signature has non zero unused bits reject
7026 the signature. All current signature algorithms require zero unused bits.
7028 2. Check certificate algorithm consistency.
7030 Check the AlgorithmIdentifier inside TBS matches the one in the
7031 certificate signature. NB: this will result in signature failure
7032 errors for some broken certificates.
7034 Thanks to Konrad Kraszewski from Google for reporting this issue.
7036 3. Check DSA/ECDSA signatures use DER.
7038 Re-encode DSA/ECDSA signatures and compare with the original received
7039 signature. Return an error if there is a mismatch.
7041 This will reject various cases including garbage after signature
7042 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7043 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7044 (negative or with leading zeroes).
7046 Further analysis was conducted and fixes were developed by Stephen Henson
7047 of the OpenSSL core team.
7053 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7054 results on some platforms, including x86_64. This bug occurs at random
7055 with a very low probability, and is not known to be exploitable in any
7056 way, though its exact impact is difficult to determine. Thanks to Pieter
7057 Wuille (Blockstream) who reported this issue and also suggested an initial
7058 fix. Further analysis was conducted by the OpenSSL development team and
7059 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7060 the OpenSSL core team.
7065 * Do not resume sessions on the server if the negotiated protocol
7066 version does not match the session's version. Resuming with a different
7067 version, while not strictly forbidden by the RFC, is of questionable
7068 sanity and breaks all known clients.
7070 *David Benjamin, Emilia Käsper*
7072 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7073 early CCS messages during renegotiation. (Note that because
7074 renegotiation is encrypted, this early CCS was not exploitable.)
7078 * Tighten client-side session ticket handling during renegotiation:
7079 ensure that the client only accepts a session ticket if the server sends
7080 the extension anew in the ServerHello. Previously, a TLS client would
7081 reuse the old extension state and thus accept a session ticket if one was
7082 announced in the initial ServerHello.
7084 Similarly, ensure that the client requires a session ticket if one
7085 was advertised in the ServerHello. Previously, a TLS client would
7086 ignore a missing NewSessionTicket message.
7090 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7094 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7095 sends a carefully crafted handshake message, to cause OpenSSL to fail
7096 to free up to 64k of memory causing a memory leak. This could be
7097 exploited in a Denial Of Service attack. This issue affects OpenSSL
7098 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7099 whether SRTP is used or configured. Implementations of OpenSSL that
7100 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7102 The fix was developed by the OpenSSL team.
7107 * Session Ticket Memory Leak.
7109 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7110 integrity of that ticket is first verified. In the event of a session
7111 ticket integrity check failing, OpenSSL will fail to free memory
7112 causing a memory leak. By sending a large number of invalid session
7113 tickets an attacker could exploit this issue in a Denial Of Service
7119 * Build option no-ssl3 is incomplete.
7121 When OpenSSL is configured with "no-ssl3" as a build option, servers
7122 could accept and complete a SSL 3.0 handshake, and clients could be
7123 configured to send them.
7126 *Akamai and the OpenSSL team*
7128 * Add support for TLS_FALLBACK_SCSV.
7129 Client applications doing fallback retries should call
7130 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7133 *Adam Langley, Bodo Moeller*
7135 * Add additional DigestInfo checks.
7137 Re-encode DigestInto in DER and check against the original when
7138 verifying RSA signature: this will reject any improperly encoded
7139 DigestInfo structures.
7141 Note: this is a precautionary measure and no attacks are currently known.
7145 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7147 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7148 SRP code can be overrun an internal buffer. Add sanity check that
7149 g, A, B < N to SRP code.
7151 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7152 Group for discovering this issue.
7157 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7158 TLS 1.0 instead of higher protocol versions when the ClientHello message
7159 is badly fragmented. This allows a man-in-the-middle attacker to force a
7160 downgrade to TLS 1.0 even if both the server and the client support a
7161 higher protocol version, by modifying the client's TLS records.
7163 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7164 researching this issue.
7169 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7170 to a denial of service attack. A malicious server can crash the client
7171 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7172 ciphersuite and sending carefully crafted handshake messages.
7174 Thanks to Felix Gröbert (Google) for discovering and researching this
7180 * By sending carefully crafted DTLS packets an attacker could cause openssl
7181 to leak memory. This can be exploited through a Denial of Service attack.
7182 Thanks to Adam Langley for discovering and researching this issue.
7187 * An attacker can force openssl to consume large amounts of memory whilst
7188 processing DTLS handshake messages. This can be exploited through a
7189 Denial of Service attack.
7190 Thanks to Adam Langley for discovering and researching this issue.
7195 * An attacker can force an error condition which causes openssl to crash
7196 whilst processing DTLS packets due to memory being freed twice. This
7197 can be exploited through a Denial of Service attack.
7198 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7204 * If a multithreaded client connects to a malicious server using a resumed
7205 session and the server sends an ec point format extension it could write
7206 up to 255 bytes to freed memory.
7208 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7214 * A malicious server can crash an OpenSSL client with a null pointer
7215 dereference (read) by specifying an SRP ciphersuite even though it was not
7216 properly negotiated with the client. This can be exploited through a
7217 Denial of Service attack.
7219 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7220 discovering and researching this issue.
7225 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7226 X509_name_oneline, X509_name_print_ex et al. to leak some information
7227 from the stack. Applications may be affected if they echo pretty printing
7228 output to the attacker.
7230 Thanks to Ivan Fratric (Google) for discovering this issue.
7233 *Emilia Käsper, and Steve Henson*
7235 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7236 for corner cases. (Certain input points at infinity could lead to
7237 bogus results, with non-infinity inputs mapped to infinity too.)
7241 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7243 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7244 handshake can force the use of weak keying material in OpenSSL
7245 SSL/TLS clients and servers.
7247 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7248 researching this issue. ([CVE-2014-0224])
7250 *KIKUCHI Masashi, Steve Henson*
7252 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7253 OpenSSL DTLS client the code can be made to recurse eventually crashing
7256 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7259 *Imre Rad, Steve Henson*
7261 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7262 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7263 client or server. This is potentially exploitable to run arbitrary
7264 code on a vulnerable client or server.
7266 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7268 *Jüri Aedla, Steve Henson*
7270 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7271 are subject to a denial of service attack.
7273 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7274 this issue. ([CVE-2014-3470])
7276 *Felix Gröbert, Ivan Fratric, Steve Henson*
7278 * Harmonize version and its documentation. -f flag is used to display
7281 *mancha <mancha1@zoho.com>*
7283 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7284 in i2d_ECPrivateKey.
7286 *mancha <mancha1@zoho.com>*
7288 * Fix some double frees. These are not thought to be exploitable.
7290 *mancha <mancha1@zoho.com>*
7292 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7294 * A missing bounds check in the handling of the TLS heartbeat extension
7295 can be used to reveal up to 64k of memory to a connected client or
7298 Thanks for Neel Mehta of Google Security for discovering this bug and to
7299 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7300 preparing the fix ([CVE-2014-0160])
7302 *Adam Langley, Bodo Moeller*
7304 * Fix for the attack described in the paper "Recovering OpenSSL
7305 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7306 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7307 <http://eprint.iacr.org/2014/140>
7309 Thanks to Yuval Yarom and Naomi Benger for discovering this
7310 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7312 *Yuval Yarom and Naomi Benger*
7314 * TLS pad extension: draft-agl-tls-padding-03
7316 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7317 TLS client Hello record length value would otherwise be > 255 and
7318 less that 512 pad with a dummy extension containing zeroes so it
7319 is at least 512 bytes long.
7321 *Adam Langley, Steve Henson*
7323 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7325 * Fix for TLS record tampering bug. A carefully crafted invalid
7326 handshake could crash OpenSSL with a NULL pointer exception.
7327 Thanks to Anton Johansson for reporting this issues.
7330 * Keep original DTLS digest and encryption contexts in retransmission
7331 structures so we can use the previous session parameters if they need
7332 to be resent. ([CVE-2013-6450])
7336 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7337 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7338 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7339 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7340 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7341 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7343 *Rob Stradling, Adam Langley*
7345 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7347 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7348 supporting platforms or when small records were transferred.
7350 *Andy Polyakov, Steve Henson*
7352 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7354 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7356 This addresses the flaw in CBC record processing discovered by
7357 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7358 at: <http://www.isg.rhul.ac.uk/tls/>
7360 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7361 Security Group at Royal Holloway, University of London
7362 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7363 Emilia Käsper for the initial patch.
7366 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7368 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7369 ciphersuites which can be exploited in a denial of service attack.
7370 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7371 and detecting this bug and to Wolfgang Ettlinger
7372 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7377 * Return an error when checking OCSP signatures when key is NULL.
7378 This fixes a DoS attack. ([CVE-2013-0166])
7382 * Make openssl verify return errors.
7384 *Chris Palmer <palmer@google.com> and Ben Laurie*
7386 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7387 the right response is stapled. Also change SSL_get_certificate()
7388 so it returns the certificate actually sent.
7389 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7391 *Rob Stradling <rob.stradling@comodo.com>*
7393 * Fix possible deadlock when decoding public keys.
7397 * Don't use TLS 1.0 record version number in initial client hello
7402 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7404 * Sanity check record length before skipping explicit IV in TLS
7405 1.2, 1.1 and DTLS to fix DoS attack.
7407 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7408 fuzzing as a service testing platform.
7413 * Initialise tkeylen properly when encrypting CMS messages.
7414 Thanks to Solar Designer of Openwall for reporting this issue.
7418 * In FIPS mode don't try to use composite ciphers as they are not
7423 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7425 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7426 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7427 mean any application compiled against OpenSSL 1.0.0 headers setting
7428 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7429 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7430 0x10000000L Any application which was previously compiled against
7431 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7432 will need to be recompiled as a result. Letting be results in
7433 inability to disable specifically TLS 1.1 and in client context,
7434 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7438 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7439 disable just protocol X, but all protocols above X *if* there are
7440 protocols *below* X still enabled. In more practical terms it means
7441 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7442 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7443 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7448 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7450 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7451 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7452 in CRYPTO_realloc_clean.
7454 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7455 issue and to Adam Langley <agl@chromium.org> for fixing it.
7458 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7460 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7464 * Workarounds for some broken servers that "hang" if a client hello
7465 record length exceeds 255 bytes.
7467 1. Do not use record version number > TLS 1.0 in initial client
7468 hello: some (but not all) hanging servers will now work.
7469 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7470 the number of ciphers sent in the client hello. This should be
7471 set to an even number, such as 50, for example by passing:
7472 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7473 Most broken servers should now work.
7474 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7475 TLS 1.2 client support entirely.
7479 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7483 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7485 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7486 STRING form instead of a DigestInfo.
7490 * The format used for MDC2 RSA signatures is inconsistent between EVP
7491 and the RSA_sign/RSA_verify functions. This was made more apparent when
7492 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7493 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7494 the correct format in RSA_verify so both forms transparently work.
7498 * Some servers which support TLS 1.0 can choke if we initially indicate
7499 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7500 encrypted premaster secret. As a workaround use the maximum permitted
7501 client version in client hello, this should keep such servers happy
7502 and still work with previous versions of OpenSSL.
7506 * Add support for TLS/DTLS heartbeats.
7508 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7510 * Add support for SCTP.
7512 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7514 * Improved PRNG seeding for VOS.
7516 *Paul Green <Paul.Green@stratus.com>*
7518 * Extensive assembler packs updates, most notably:
7520 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7521 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7522 - x86_64: bit-sliced AES implementation;
7523 - ARM: NEON support, contemporary platforms optimizations;
7524 - s390x: z196 support;
7525 - `*`: GHASH and GF(2^m) multiplication implementations;
7529 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7530 (removal of unnecessary code)
7532 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7534 * Add TLS key material exporter from RFC 5705.
7538 * Add DTLS-SRTP negotiation from RFC 5764.
7542 * Add Next Protocol Negotiation,
7543 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7544 disabled with a no-npn flag to config or Configure. Code donated
7547 *Adam Langley <agl@google.com> and Ben Laurie*
7549 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7550 NIST-P256, NIST-P521, with constant-time single point multiplication on
7551 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7552 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7553 Code made available under Apache License version 2.0.
7555 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7556 line to include this in your build of OpenSSL, and run "make depend" (or
7557 "make update"). This enables the following EC_METHODs:
7559 EC_GFp_nistp224_method()
7560 EC_GFp_nistp256_method()
7561 EC_GFp_nistp521_method()
7563 EC_GROUP_new_by_curve_name() will automatically use these (while
7564 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7567 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7569 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7570 all platforms. Move ssize_t definition from e_os.h to the public
7571 header file e_os2.h as it now appears in public header file cms.h
7575 * New -sigopt option to the ca, req and x509 utilities. Additional
7576 signature parameters can be passed using this option and in
7581 * Add RSA PSS signing function. This will generate and set the
7582 appropriate AlgorithmIdentifiers for PSS based on those in the
7583 corresponding EVP_MD_CTX structure. No application support yet.
7587 * Support for companion algorithm specific ASN1 signing routines.
7588 New function ASN1_item_sign_ctx() signs a pre-initialised
7589 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7590 the appropriate parameters.
7594 * Add new algorithm specific ASN1 verification initialisation function
7595 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7596 handling will be the same no matter what EVP_PKEY_METHOD is used.
7597 Add a PSS handler to support verification of PSS signatures: checked
7598 against a number of sample certificates.
7602 * Add signature printing for PSS. Add PSS OIDs.
7604 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7606 * Add algorithm specific signature printing. An individual ASN1 method
7607 can now print out signatures instead of the standard hex dump.
7609 More complex signatures (e.g. PSS) can print out more meaningful
7610 information. Include DSA version that prints out the signature
7615 * Password based recipient info support for CMS library: implementing
7620 * Split password based encryption into PBES2 and PBKDF2 functions. This
7621 neatly separates the code into cipher and PBE sections and is required
7622 for some algorithms that split PBES2 into separate pieces (such as
7623 password based CMS).
7627 * Session-handling fixes:
7628 - Fix handling of connections that are resuming with a session ID,
7629 but also support Session Tickets.
7630 - Fix a bug that suppressed issuing of a new ticket if the client
7631 presented a ticket with an expired session.
7632 - Try to set the ticket lifetime hint to something reasonable.
7633 - Make tickets shorter by excluding irrelevant information.
7634 - On the client side, don't ignore renewed tickets.
7636 *Adam Langley, Bodo Moeller (Google)*
7638 * Fix PSK session representation.
7642 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7644 This work was sponsored by Intel.
7648 * Add GCM support to TLS library. Some custom code is needed to split
7649 the IV between the fixed (from PRF) and explicit (from TLS record)
7650 portions. This adds all GCM ciphersuites supported by RFC5288 and
7651 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7652 add a special AESGCM string for GCM only.
7656 * Expand range of ctrls for AES GCM. Permit setting invocation
7657 field on decrypt and retrieval of invocation field only on encrypt.
7661 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7662 As required by RFC5289 these ciphersuites cannot be used if for
7663 versions of TLS earlier than 1.2.
7667 * For FIPS capable OpenSSL interpret a NULL default public key method
7668 as unset and return the appropriate default but do *not* set the default.
7669 This means we can return the appropriate method in applications that
7670 switch between FIPS and non-FIPS modes.
7674 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7675 ENGINE is used then we cannot handle that in the FIPS module so we
7676 keep original code iff non-FIPS operations are allowed.
7680 * Add -attime option to openssl utilities.
7682 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7684 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7688 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7689 FIPS EC methods unconditionally for now.
7693 * New build option no-ec2m to disable characteristic 2 code.
7697 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7698 all cases can be covered as some introduce binary incompatibilities.
7702 * Redirect RSA operations to FIPS module including keygen,
7703 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7707 * Add similar low-level API blocking to ciphers.
7711 * low-level digest APIs are not approved in FIPS mode: any attempt
7712 to use these will cause a fatal error. Applications that *really* want
7713 to use them can use the `private_*` version instead.
7717 * Redirect cipher operations to FIPS module for FIPS builds.
7721 * Redirect digest operations to FIPS module for FIPS builds.
7725 * Update build system to add "fips" flag which will link in fipscanister.o
7726 for static and shared library builds embedding a signature if needed.
7730 * Output TLS supported curves in preference order instead of numerical
7731 order. This is currently hardcoded for the highest order curves first.
7732 This should be configurable so applications can judge speed vs strength.
7736 * Add TLS v1.2 server support for client authentication.
7740 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7745 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7746 FIPS modules versions.
7750 * Add TLS v1.2 client side support for client authentication. Keep cache
7751 of handshake records longer as we don't know the hash algorithm to use
7752 until after the certificate request message is received.
7756 * Initial TLS v1.2 client support. Add a default signature algorithms
7757 extension including all the algorithms we support. Parse new signature
7758 format in client key exchange. Relax some ECC signing restrictions for
7759 TLS v1.2 as indicated in RFC5246.
7763 * Add server support for TLS v1.2 signature algorithms extension. Switch
7764 to new signature format when needed using client digest preference.
7765 All server ciphersuites should now work correctly in TLS v1.2. No client
7766 support yet and no support for client certificates.
7770 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7771 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7772 ciphersuites. At present only RSA key exchange ciphersuites work with
7773 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7774 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7775 and version checking.
7779 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7780 with this defined it will not be affected by any changes to ssl internal
7781 structures. Add several utility functions to allow openssl application
7782 to work with OPENSSL_NO_SSL_INTERN defined.
7786 * A long standing patch to add support for SRP from EdelWeb (Peter
7787 Sylvester and Christophe Renou) was integrated.
7788 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7789 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7792 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7796 * Permit abbreviated handshakes when renegotiating using the function
7797 SSL_renegotiate_abbreviated().
7799 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7801 * Add call to ENGINE_register_all_complete() to
7802 ENGINE_load_builtin_engines(), so some implementations get used
7803 automatically instead of needing explicit application support.
7807 * Add support for TLS key exporter as described in RFC5705.
7809 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7811 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7812 a few changes are required:
7814 Add SSL_OP_NO_TLSv1_1 flag.
7815 Add TLSv1_1 methods.
7816 Update version checking logic to handle version 1.1.
7817 Add explicit IV handling (ported from DTLS code).
7818 Add command line options to s_client/s_server.
7825 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7827 * X509_ATTRIBUTE memory leak
7829 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7830 memory. This structure is used by the PKCS#7 and CMS routines so any
7831 application which reads PKCS#7 or CMS data from untrusted sources is
7832 affected. SSL/TLS is not affected.
7834 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7840 * Race condition handling PSK identify hint
7842 If PSK identity hints are received by a multi-threaded client then
7843 the values are wrongly updated in the parent SSL_CTX structure. This can
7844 result in a race condition potentially leading to a double free of the
7850 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7852 * Malformed ECParameters causes infinite loop
7854 When processing an ECParameters structure OpenSSL enters an infinite loop
7855 if the curve specified is over a specially malformed binary polynomial
7858 This can be used to perform denial of service against any
7859 system which processes public keys, certificate requests or
7860 certificates. This includes TLS clients and TLS servers with
7861 client authentication enabled.
7863 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7868 * Exploitable out-of-bounds read in X509_cmp_time
7870 X509_cmp_time does not properly check the length of the ASN1_TIME
7871 string and can read a few bytes out of bounds. In addition,
7872 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7875 An attacker can use this to craft malformed certificates and CRLs of
7876 various sizes and potentially cause a segmentation fault, resulting in
7877 a DoS on applications that verify certificates or CRLs. TLS clients
7878 that verify CRLs are affected. TLS clients and servers with client
7879 authentication enabled may be affected if they use custom verification
7882 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7883 independently by Hanno Böck.
7888 * PKCS7 crash with missing EnvelopedContent
7890 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7891 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7892 with missing content and trigger a NULL pointer dereference on parsing.
7894 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7895 structures from untrusted sources are affected. OpenSSL clients and
7896 servers are not affected.
7898 This issue was reported to OpenSSL by Michal Zalewski (Google).
7903 * CMS verify infinite loop with unknown hash function
7905 When verifying a signedData message the CMS code can enter an infinite loop
7906 if presented with an unknown hash function OID. This can be used to perform
7907 denial of service against any system which verifies signedData messages using
7909 This issue was reported to OpenSSL by Johannes Bauer.
7914 * Race condition handling NewSessionTicket
7916 If a NewSessionTicket is received by a multi-threaded client when attempting to
7917 reuse a previous ticket then a race condition can occur potentially leading to
7918 a double free of the ticket data.
7923 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7925 * Segmentation fault in ASN1_TYPE_cmp fix
7927 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7928 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7929 certificate signature algorithm consistency this can be used to crash any
7930 certificate verification operation and exploited in a DoS attack. Any
7931 application which performs certificate verification is vulnerable including
7932 OpenSSL clients and servers which enable client authentication.
7937 * ASN.1 structure reuse memory corruption fix
7939 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7940 memory corruption via an invalid write. Such reuse is and has been
7941 strongly discouraged and is believed to be rare.
7943 Applications that parse structures containing CHOICE or ANY DEFINED BY
7944 components may be affected. Certificate parsing (d2i_X509 and related
7945 functions) are however not affected. OpenSSL clients and servers are
7951 * PKCS7 NULL pointer dereferences fix
7953 The PKCS#7 parsing code does not handle missing outer ContentInfo
7954 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7955 missing content and trigger a NULL pointer dereference on parsing.
7957 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7958 otherwise parse PKCS#7 structures from untrusted sources are
7959 affected. OpenSSL clients and servers are not affected.
7961 This issue was reported to OpenSSL by Michal Zalewski (Google).
7966 * DoS via reachable assert in SSLv2 servers fix
7968 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7969 servers that both support SSLv2 and enable export cipher suites by sending
7970 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7972 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7973 (OpenSSL development team).
7978 * Use After Free following d2i_ECPrivatekey error fix
7980 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7981 could cause a use after free condition. This, in turn, could cause a double
7982 free in several private key parsing functions (such as d2i_PrivateKey
7983 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7984 for applications that receive EC private keys from untrusted
7985 sources. This scenario is considered rare.
7987 This issue was discovered by the BoringSSL project and fixed in their
7993 * X509_to_X509_REQ NULL pointer deref fix
7995 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7996 the certificate key is invalid. This function is rarely used in practice.
7998 This issue was discovered by Brian Carpenter.
8003 * Removed the export ciphers from the DEFAULT ciphers
8007 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8009 * Build fixes for the Windows and OpenVMS platforms
8011 *Matt Caswell and Richard Levitte*
8013 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8015 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8016 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8017 dereference. This could lead to a Denial Of Service attack. Thanks to
8018 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8023 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8024 dtls1_buffer_record function under certain conditions. In particular this
8025 could occur if an attacker sent repeated DTLS records with the same
8026 sequence number but for the next epoch. The memory leak could be exploited
8027 by an attacker in a Denial of Service attack through memory exhaustion.
8028 Thanks to Chris Mueller for reporting this issue.
8033 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8034 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8035 method would be set to NULL which could later result in a NULL pointer
8036 dereference. Thanks to Frank Schmirler for reporting this issue.
8041 * Abort handshake if server key exchange message is omitted for ephemeral
8044 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8045 reporting this issue.
8050 * Remove non-export ephemeral RSA code on client and server. This code
8051 violated the TLS standard by allowing the use of temporary RSA keys in
8052 non-export ciphersuites and could be used by a server to effectively
8053 downgrade the RSA key length used to a value smaller than the server
8054 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8055 INRIA or reporting this issue.
8060 * Fixed issue where DH client certificates are accepted without verification.
8061 An OpenSSL server will accept a DH certificate for client authentication
8062 without the certificate verify message. This effectively allows a client to
8063 authenticate without the use of a private key. This only affects servers
8064 which trust a client certificate authority which issues certificates
8065 containing DH keys: these are extremely rare and hardly ever encountered.
8066 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8072 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8073 results on some platforms, including x86_64. This bug occurs at random
8074 with a very low probability, and is not known to be exploitable in any
8075 way, though its exact impact is difficult to determine. Thanks to Pieter
8076 Wuille (Blockstream) who reported this issue and also suggested an initial
8077 fix. Further analysis was conducted by the OpenSSL development team and
8078 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8079 the OpenSSL core team.
8084 * Fix various certificate fingerprint issues.
8086 By using non-DER or invalid encodings outside the signed portion of a
8087 certificate the fingerprint can be changed without breaking the signature.
8088 Although no details of the signed portion of the certificate can be changed
8089 this can cause problems with some applications: e.g. those using the
8090 certificate fingerprint for blacklists.
8092 1. Reject signatures with non zero unused bits.
8094 If the BIT STRING containing the signature has non zero unused bits reject
8095 the signature. All current signature algorithms require zero unused bits.
8097 2. Check certificate algorithm consistency.
8099 Check the AlgorithmIdentifier inside TBS matches the one in the
8100 certificate signature. NB: this will result in signature failure
8101 errors for some broken certificates.
8103 Thanks to Konrad Kraszewski from Google for reporting this issue.
8105 3. Check DSA/ECDSA signatures use DER.
8107 Reencode DSA/ECDSA signatures and compare with the original received
8108 signature. Return an error if there is a mismatch.
8110 This will reject various cases including garbage after signature
8111 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8112 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8113 (negative or with leading zeroes).
8115 Further analysis was conducted and fixes were developed by Stephen Henson
8116 of the OpenSSL core team.
8122 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8124 * Session Ticket Memory Leak.
8126 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8127 integrity of that ticket is first verified. In the event of a session
8128 ticket integrity check failing, OpenSSL will fail to free memory
8129 causing a memory leak. By sending a large number of invalid session
8130 tickets an attacker could exploit this issue in a Denial Of Service
8136 * Build option no-ssl3 is incomplete.
8138 When OpenSSL is configured with "no-ssl3" as a build option, servers
8139 could accept and complete a SSL 3.0 handshake, and clients could be
8140 configured to send them.
8143 *Akamai and the OpenSSL team*
8145 * Add support for TLS_FALLBACK_SCSV.
8146 Client applications doing fallback retries should call
8147 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8150 *Adam Langley, Bodo Moeller*
8152 * Add additional DigestInfo checks.
8154 Reencode DigestInto in DER and check against the original when
8155 verifying RSA signature: this will reject any improperly encoded
8156 DigestInfo structures.
8158 Note: this is a precautionary measure and no attacks are currently known.
8162 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8164 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8165 to a denial of service attack. A malicious server can crash the client
8166 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8167 ciphersuite and sending carefully crafted handshake messages.
8169 Thanks to Felix Gröbert (Google) for discovering and researching this
8175 * By sending carefully crafted DTLS packets an attacker could cause openssl
8176 to leak memory. This can be exploited through a Denial of Service attack.
8177 Thanks to Adam Langley for discovering and researching this issue.
8182 * An attacker can force openssl to consume large amounts of memory whilst
8183 processing DTLS handshake messages. This can be exploited through a
8184 Denial of Service attack.
8185 Thanks to Adam Langley for discovering and researching this issue.
8190 * An attacker can force an error condition which causes openssl to crash
8191 whilst processing DTLS packets due to memory being freed twice. This
8192 can be exploited through a Denial of Service attack.
8193 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8199 * If a multithreaded client connects to a malicious server using a resumed
8200 session and the server sends an ec point format extension it could write
8201 up to 255 bytes to freed memory.
8203 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8209 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8210 X509_name_oneline, X509_name_print_ex et al. to leak some information
8211 from the stack. Applications may be affected if they echo pretty printing
8212 output to the attacker.
8214 Thanks to Ivan Fratric (Google) for discovering this issue.
8217 *Emilia Käsper, and Steve Henson*
8219 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8220 for corner cases. (Certain input points at infinity could lead to
8221 bogus results, with non-infinity inputs mapped to infinity too.)
8225 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8227 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8228 handshake can force the use of weak keying material in OpenSSL
8229 SSL/TLS clients and servers.
8231 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8232 researching this issue. ([CVE-2014-0224])
8234 *KIKUCHI Masashi, Steve Henson*
8236 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8237 OpenSSL DTLS client the code can be made to recurse eventually crashing
8240 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8243 *Imre Rad, Steve Henson*
8245 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8246 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8247 client or server. This is potentially exploitable to run arbitrary
8248 code on a vulnerable client or server.
8250 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8252 *Jüri Aedla, Steve Henson*
8254 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8255 are subject to a denial of service attack.
8257 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8258 this issue. ([CVE-2014-3470])
8260 *Felix Gröbert, Ivan Fratric, Steve Henson*
8262 * Harmonize version and its documentation. -f flag is used to display
8265 *mancha <mancha1@zoho.com>*
8267 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8268 in i2d_ECPrivateKey.
8270 *mancha <mancha1@zoho.com>*
8272 * Fix some double frees. These are not thought to be exploitable.
8274 *mancha <mancha1@zoho.com>*
8276 * Fix for the attack described in the paper "Recovering OpenSSL
8277 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8278 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8279 <http://eprint.iacr.org/2014/140>
8281 Thanks to Yuval Yarom and Naomi Benger for discovering this
8282 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8284 *Yuval Yarom and Naomi Benger*
8286 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8288 * Keep original DTLS digest and encryption contexts in retransmission
8289 structures so we can use the previous session parameters if they need
8290 to be resent. ([CVE-2013-6450])
8294 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8295 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8296 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8297 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8298 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8299 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8301 *Rob Stradling, Adam Langley*
8303 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8305 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8307 This addresses the flaw in CBC record processing discovered by
8308 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8309 at: <http://www.isg.rhul.ac.uk/tls/>
8311 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8312 Security Group at Royal Holloway, University of London
8313 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8314 Emilia Käsper for the initial patch.
8317 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8319 * Return an error when checking OCSP signatures when key is NULL.
8320 This fixes a DoS attack. ([CVE-2013-0166])
8324 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8325 the right response is stapled. Also change SSL_get_certificate()
8326 so it returns the certificate actually sent.
8327 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8328 (This is a backport)
8330 *Rob Stradling <rob.stradling@comodo.com>*
8332 * Fix possible deadlock when decoding public keys.
8336 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8338 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8341 * Sanity check record length before skipping explicit IV in DTLS
8344 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8345 fuzzing as a service testing platform.
8350 * Initialise tkeylen properly when encrypting CMS messages.
8351 Thanks to Solar Designer of Openwall for reporting this issue.
8355 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8357 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8358 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8359 in CRYPTO_realloc_clean.
8361 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8362 issue and to Adam Langley <agl@chromium.org> for fixing it.
8365 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8367 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8369 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8370 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8371 content decryption and always return the same error. Note: this attack
8372 needs on average 2^20 messages so it only affects automated senders. The
8373 old behaviour can be re-enabled in the CMS code by setting the
8374 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8375 an MMA defence is not necessary.
8376 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8377 this issue. ([CVE-2012-0884])
8381 * Fix CVE-2011-4619: make sure we really are receiving a
8382 client hello before rejecting multiple SGC restarts. Thanks to
8383 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8387 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8389 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8390 Thanks to Antonio Martin, Enterprise Secure Access Research and
8391 Development, Cisco Systems, Inc. for discovering this bug and
8392 preparing a fix. ([CVE-2012-0050])
8396 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8398 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8399 of the Vaudenay padding oracle attack on CBC mode encryption
8400 which enables an efficient plaintext recovery attack against
8401 the OpenSSL implementation of DTLS. Their attack exploits timing
8402 differences arising during decryption processing. A research
8403 paper describing this attack can be found at:
8404 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8405 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8406 Security Group at Royal Holloway, University of London
8407 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8408 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8409 for preparing the fix. ([CVE-2011-4108])
8411 *Robin Seggelmann, Michael Tuexen*
8413 * Clear bytes used for block padding of SSL 3.0 records.
8416 *Adam Langley (Google)*
8418 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8419 Kadianakis <desnacked@gmail.com> for discovering this issue and
8420 Adam Langley for preparing the fix. ([CVE-2011-4619])
8422 *Adam Langley (Google)*
8424 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8426 *Andrey Kulikov <amdeich@gmail.com>*
8428 * Prevent malformed RFC3779 data triggering an assertion failure.
8429 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8430 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8432 *Rob Austein <sra@hactrn.net>*
8434 * Improved PRNG seeding for VOS.
8436 *Paul Green <Paul.Green@stratus.com>*
8438 * Fix ssl_ciph.c set-up race.
8440 *Adam Langley (Google)*
8442 * Fix spurious failures in ecdsatest.c.
8444 *Emilia Käsper (Google)*
8446 * Fix the BIO_f_buffer() implementation (which was mixing different
8447 interpretations of the `..._len` fields).
8449 *Adam Langley (Google)*
8451 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8452 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8453 threads won't reuse the same blinding coefficients.
8455 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8456 lock to call BN_BLINDING_invert_ex, and avoids one use of
8457 BN_BLINDING_update for each BN_BLINDING structure (previously,
8458 the last update always remained unused).
8460 *Emilia Käsper (Google)*
8462 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8464 *Bob Buckholz (Google)*
8466 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8468 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8469 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8471 *Kaspar Brand <ossl@velox.ch>*
8473 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8474 for multi-threaded use of ECDH. ([CVE-2011-3210])
8476 *Adam Langley (Google)*
8478 * Fix x509_name_ex_d2i memory leak on bad inputs.
8482 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8483 signature public key algorithm by using OID xref utilities instead.
8484 Before this you could only use some ECC ciphersuites with SHA1 only.
8488 * Add protection against ECDSA timing attacks as mentioned in the paper
8489 by Billy Bob Brumley and Nicola Tuveri, see:
8490 <http://eprint.iacr.org/2011/232.pdf>
8492 *Billy Bob Brumley and Nicola Tuveri*
8494 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8496 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8498 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8500 * Fix bug in string printing code: if *any* escaping is enabled we must
8501 escape the escape character (backslash) or the resulting string is
8506 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8508 * Disable code workaround for ancient and obsolete Netscape browsers
8509 and servers: an attacker can use it in a ciphersuite downgrade attack.
8510 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8514 * Fixed J-PAKE implementation error, originally discovered by
8515 Sebastien Martini, further info and confirmation from Stefan
8516 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8520 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8522 * Fix extension code to avoid race conditions which can result in a buffer
8523 overrun vulnerability: resumed sessions must not be modified as they can
8524 be shared by multiple threads. CVE-2010-3864
8528 * Fix WIN32 build system to correctly link an ENGINE directory into
8533 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8535 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8538 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8540 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8542 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8543 context. The operation can be customised via the ctrl mechanism in
8544 case ENGINEs want to include additional functionality.
8548 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8552 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8553 output hashes compatible with older versions of OpenSSL.
8555 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8557 * Fix compression algorithm handling: if resuming a session use the
8558 compression algorithm of the resumed session instead of determining
8559 it from client hello again. Don't allow server to change algorithm.
8563 * Add load_crls() function to commands tidying load_certs() too. Add option
8564 to verify utility to allow additional CRLs to be included.
8568 * Update OCSP request code to permit adding custom headers to the request:
8569 some responders need this.
8573 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8576 *Julia Lawall <julia@diku.dk>*
8578 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8579 needlessly dereferenced structures, used obsolete functions and
8580 didn't handle all updated verify codes correctly.
8584 * Disable MD2 in the default configuration.
8588 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8589 indicate the initial BIO being pushed or popped. This makes it possible
8590 to determine whether the BIO is the one explicitly called or as a result
8591 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8592 it handles reference counts correctly and doesn't zero out the I/O bio
8593 when it is not being explicitly popped. WARNING: applications which
8594 included workarounds for the old buggy behaviour will need to be modified
8595 or they could free up already freed BIOs.
8599 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8600 renaming to all platforms (within the 0.9.8 branch, this was
8601 done conditionally on Netware platforms to avoid a name clash).
8603 *Guenter <lists@gknw.net>*
8605 * Add ECDHE and PSK support to DTLS.
8607 *Michael Tuexen <tuexen@fh-muenster.de>*
8609 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8614 * Add "missing" function EVP_MD_flags() (without this the only way to
8615 retrieve a digest flags is by accessing the structure directly. Update
8616 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8617 or cipher is registered as in the "from" argument. Print out all
8618 registered digests in the dgst usage message instead of manually
8619 attempting to work them out.
8623 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8624 this allows the use of compression and extensions. Change default cipher
8625 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8626 by default unless an application cipher string requests it.
8630 * Alter match criteria in PKCS12_parse(). It used to try to use local
8631 key ids to find matching certificates and keys but some PKCS#12 files
8632 don't follow the (somewhat unwritten) rules and this strategy fails.
8633 Now just gather all certificates together and the first private key
8634 then look for the first certificate that matches the key.
8638 * Support use of registered digest and cipher names for dgst and cipher
8639 commands instead of having to add each one as a special case. So now
8646 openssl dgst -sha256 foo
8648 and this works for ENGINE based algorithms too.
8652 * Update Gost ENGINE to support parameter files.
8654 *Victor B. Wagner <vitus@cryptocom.ru>*
8656 * Support GeneralizedTime in ca utility.
8658 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8660 * Enhance the hash format used for certificate directory links. The new
8661 form uses the canonical encoding (meaning equivalent names will work
8662 even if they aren't identical) and uses SHA1 instead of MD5. This form
8663 is incompatible with the older format and as a result c_rehash should
8664 be used to rebuild symbolic links.
8668 * Make PKCS#8 the default write format for private keys, replacing the
8669 traditional format. This form is standardised, more secure and doesn't
8670 include an implicit MD5 dependency.
8674 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8675 committed to OpenSSL should pass this lot as a minimum.
8679 * Add session ticket override functionality for use by EAP-FAST.
8681 *Jouni Malinen <j@w1.fi>*
8683 * Modify HMAC functions to return a value. Since these can be implemented
8684 in an ENGINE errors can occur.
8688 * Type-checked OBJ_bsearch_ex.
8692 * Type-checked OBJ_bsearch. Also some constification necessitated
8693 by type-checking. Still to come: TXT_DB, bsearch(?),
8694 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8699 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8700 seconds to a tm structure directly, instead of going through OS
8701 specific date routines. This avoids any issues with OS routines such
8702 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8703 and X509_time_adj_ex() to cover the extended range. The existing
8704 X509_time_adj() is still usable and will no longer have any date issues.
8708 * Delta CRL support. New use deltas option which will attempt to locate
8709 and search any appropriate delta CRLs available.
8711 This work was sponsored by Google.
8715 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8716 code and add additional score elements. Validate alternate CRL paths
8717 as part of the CRL checking and indicate a new error "CRL path validation
8718 error" in this case. Applications wanting additional details can use
8719 the verify callback and check the new "parent" field. If this is not
8720 NULL CRL path validation is taking place. Existing applications won't
8721 see this because it requires extended CRL support which is off by
8724 This work was sponsored by Google.
8728 * Support for freshest CRL extension.
8730 This work was sponsored by Google.
8734 * Initial indirect CRL support. Currently only supported in the CRLs
8735 passed directly and not via lookup. Process certificate issuer
8736 CRL entry extension and lookup CRL entries by bother issuer name
8737 and serial number. Check and process CRL issuer entry in IDP extension.
8739 This work was sponsored by Google.
8743 * Add support for distinct certificate and CRL paths. The CRL issuer
8744 certificate is validated separately in this case. Only enabled if
8745 an extended CRL support flag is set: this flag will enable additional
8746 CRL functionality in future.
8748 This work was sponsored by Google.
8752 * Add support for policy mappings extension.
8754 This work was sponsored by Google.
8758 * Fixes to pathlength constraint, self issued certificate handling,
8759 policy processing to align with RFC3280 and PKITS tests.
8761 This work was sponsored by Google.
8765 * Support for name constraints certificate extension. DN, email, DNS
8766 and URI types are currently supported.
8768 This work was sponsored by Google.
8772 * To cater for systems that provide a pointer-based thread ID rather
8773 than numeric, deprecate the current numeric thread ID mechanism and
8774 replace it with a structure and associated callback type. This
8775 mechanism allows a numeric "hash" to be extracted from a thread ID in
8776 either case, and on platforms where pointers are larger than 'long',
8777 mixing is done to help ensure the numeric 'hash' is usable even if it
8778 can't be guaranteed unique. The default mechanism is to use "&errno"
8779 as a pointer-based thread ID to distinguish between threads.
8781 Applications that want to provide their own thread IDs should now use
8782 CRYPTO_THREADID_set_callback() to register a callback that will call
8783 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8785 Note that ERR_remove_state() is now deprecated, because it is tied
8786 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8787 to free the current thread's error state should be replaced by
8788 ERR_remove_thread_state(NULL).
8790 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8791 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8792 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8793 application was previously providing a numeric thread callback that
8794 was inappropriate for distinguishing threads, then uniqueness might
8795 have been obtained with &errno that happened immediately in the
8796 intermediate development versions of OpenSSL; this is no longer the
8797 case, the numeric thread callback will now override the automatic use
8800 *Geoff Thorpe, with help from Bodo Moeller*
8802 * Initial support for different CRL issuing certificates. This covers a
8803 simple case where the self issued certificates in the chain exist and
8804 the real CRL issuer is higher in the existing chain.
8806 This work was sponsored by Google.
8810 * Removed effectively defunct crypto/store from the build.
8814 * Revamp of STACK to provide stronger type-checking. Still to come:
8815 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8816 ASN1_STRING, CONF_VALUE.
8820 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8821 RAM on SSL connections. This option can save about 34k per idle SSL.
8825 * Revamp of LHASH to provide stronger type-checking. Still to come:
8826 STACK, TXT_DB, bsearch, qsort.
8830 * Initial support for Cryptographic Message Syntax (aka CMS) based
8831 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8832 support for data, signedData, compressedData, digestedData and
8833 encryptedData, envelopedData types included. Scripts to check against
8834 RFC4134 examples draft and interop and consistency checks of many
8835 content types and variants.
8839 * Add options to enc utility to support use of zlib compression BIO.
8843 * Extend mk1mf to support importing of options and assembly language
8844 files from Configure script, currently only included in VC-WIN32.
8845 The assembly language rules can now optionally generate the source
8846 files from the associated perl scripts.
8850 * Implement remaining functionality needed to support GOST ciphersuites.
8851 Interop testing has been performed using CryptoPro implementations.
8853 *Victor B. Wagner <vitus@cryptocom.ru>*
8855 * s390x assembler pack.
8859 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8864 * Implement Opaque PRF Input TLS extension as specified in
8865 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8866 official specification yet and no extension type assignment by
8867 IANA exists, this extension (for now) will have to be explicitly
8868 enabled when building OpenSSL by providing the extension number
8869 to use. For example, specify an option
8871 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8873 to the "config" or "Configure" script to enable the extension,
8874 assuming extension number 0x9527 (which is a completely arbitrary
8875 and unofficial assignment based on the MD5 hash of the Internet
8876 Draft). Note that by doing so, you potentially lose
8877 interoperability with other TLS implementations since these might
8878 be using the same extension number for other purposes.
8880 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8881 opaque PRF input value to use in the handshake. This will create
8882 an internal copy of the length-'len' string at 'src', and will
8883 return non-zero for success.
8885 To get more control and flexibility, provide a callback function
8888 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8889 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8893 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8896 Callback function 'cb' will be called in handshakes, and is
8897 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8898 Argument 'arg' is for application purposes (the value as given to
8899 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8900 be provided to the callback function). The callback function
8901 has to return non-zero to report success: usually 1 to use opaque
8902 PRF input just if possible, or 2 to enforce use of the opaque PRF
8903 input. In the latter case, the library will abort the handshake
8904 if opaque PRF input is not successfully negotiated.
8906 Arguments 'peerinput' and 'len' given to the callback function
8907 will always be NULL and 0 in the case of a client. A server will
8908 see the client's opaque PRF input through these variables if
8909 available (NULL and 0 otherwise). Note that if the server
8910 provides an opaque PRF input, the length must be the same as the
8911 length of the client's opaque PRF input.
8913 Note that the callback function will only be called when creating
8914 a new session (session resumption can resume whatever was
8915 previously negotiated), and will not be called in SSL 2.0
8916 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8917 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8918 for applications that need to enforce opaque PRF input.
8922 * Update ssl code to support digests other than SHA1+MD5 for handshake
8925 *Victor B. Wagner <vitus@cryptocom.ru>*
8927 * Add RFC4507 support to OpenSSL. This includes the corrections in
8928 RFC4507bis. The encrypted ticket format is an encrypted encoded
8929 SSL_SESSION structure, that way new session features are automatically
8932 If a client application caches session in an SSL_SESSION structure
8933 support is transparent because tickets are now stored in the encoded
8936 The SSL_CTX structure automatically generates keys for ticket
8937 protection in servers so again support should be possible
8938 with no application modification.
8940 If a client or server wishes to disable RFC4507 support then the option
8941 SSL_OP_NO_TICKET can be set.
8943 Add a TLS extension debugging callback to allow the contents of any client
8944 or server extensions to be examined.
8946 This work was sponsored by Google.
8950 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8951 OpenSSL should now compile cleanly on gcc 4.2
8953 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8955 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8956 support including streaming MAC support: this is required for GOST
8957 ciphersuite support.
8959 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8961 * Add option -stream to use PKCS#7 streaming in smime utility. New
8962 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8963 to output in BER and PEM format.
8967 * Experimental support for use of HMAC via EVP_PKEY interface. This
8968 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8969 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8970 ENGINE support for HMAC keys which are unextractable. New -mac and
8971 -macopt options to dgst utility.
8975 * New option -sigopt to dgst utility. Update dgst to use
8976 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8977 alternative signing parameters such as X9.31 or PSS in the dgst
8982 * Change ssl_cipher_apply_rule(), the internal function that does
8983 the work each time a ciphersuite string requests enabling
8984 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8985 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8986 the order of disabled ciphersuites such that those ciphersuites
8987 that most recently went from enabled to disabled not only stay
8988 in order with respect to each other, but also have higher priority
8989 than other disabled ciphersuites the next time ciphersuites are
8992 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8993 the same ciphersuites as with "HIGH" alone, but in a specific
8994 order where the PSK ciphersuites come first (since they are the
8995 most recently disabled ciphersuites when "HIGH" is parsed).
8997 Also, change ssl_create_cipher_list() (using this new
8998 functionality) such that between otherwise identical
8999 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9004 * Change ssl_create_cipher_list() so that it automatically
9005 arranges the ciphersuites in reasonable order before starting
9006 to process the rule string. Thus, the definition for "DEFAULT"
9007 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9008 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9009 This makes it much easier to arrive at a reasonable default order
9010 in applications for which anonymous ciphers are OK (meaning
9011 that you can't actually use DEFAULT).
9013 *Bodo Moeller; suggested by Victor Duchovni*
9015 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9016 processing) into multiple integers instead of setting
9017 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9018 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9019 (These masks as well as the individual bit definitions are hidden
9020 away into the non-exported interface ssl/ssl_locl.h, so this
9021 change to the definition of the SSL_CIPHER structure shouldn't
9022 affect applications.) This give us more bits for each of these
9023 categories, so there is no longer a need to coagulate AES128 and
9024 AES256 into a single algorithm bit, and to coagulate Camellia128
9025 and Camellia256 into a single algorithm bit, which has led to all
9028 Thus, among other things, the kludge introduced in 0.9.7m and
9029 0.9.8e for masking out AES256 independently of AES128 or masking
9030 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9032 With the change, we also introduce new ciphersuite aliases that
9033 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9038 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9039 Use the leftmost N bytes of the signature input if the input is
9040 larger than the prime q (with N being the size in bytes of q).
9044 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9045 it yet and it is largely untested.
9049 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9053 * Initial incomplete changes to avoid need for function casts in OpenSSL
9054 some compilers (gcc 4.2 and later) reject their use. Safestack is
9055 reimplemented. Update ASN1 to avoid use of legacy functions.
9059 * Win32/64 targets are linked with Winsock2.
9063 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9064 to external functions. This can be used to increase CRL handling
9065 efficiency especially when CRLs are very large by (for example) storing
9066 the CRL revoked certificates in a database.
9070 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9071 new CRLs added to a directory can be used. New command line option
9072 -verify_return_error to s_client and s_server. This causes real errors
9073 to be returned by the verify callback instead of carrying on no matter
9074 what. This reflects the way a "real world" verify callback would behave.
9078 * GOST engine, supporting several GOST algorithms and public key formats.
9079 Kindly donated by Cryptocom.
9083 * Partial support for Issuing Distribution Point CRL extension. CRLs
9084 partitioned by DP are handled but no indirect CRL or reason partitioning
9085 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9086 selected via a scoring technique which handles IDP and AKID in CRLs.
9090 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9091 will ultimately be used for all verify operations: this will remove the
9092 X509_STORE dependency on certificate verification and allow alternative
9093 lookup methods. X509_STORE based implementations of these two callbacks.
9097 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9098 Modify get_crl() to find a valid (unexpired) CRL if possible.
9102 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9103 this would be called X509_CRL_cmp() but that name is already used by
9104 a function that just compares CRL issuer names. Cache several CRL
9105 extensions in X509_CRL structure and cache CRLDP in X509.
9109 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9110 this maps equivalent X509_NAME structures into a consistent structure.
9111 Name comparison can then be performed rapidly using memcmp().
9115 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9120 * Allow digests to supply their own micalg string for S/MIME type using
9121 the ctrl EVP_MD_CTRL_MICALG.
9125 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9126 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9127 ctrl. It can then customise the structure before and/or after signing
9132 * New function OBJ_add_sigid() to allow application defined signature OIDs
9133 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9134 to free up any added signature OIDs.
9138 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9139 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9140 digest and cipher tables. New options added to openssl utility:
9141 list-message-digest-algorithms and list-cipher-algorithms.
9145 * Change the array representation of binary polynomials: the list
9146 of degrees of non-zero coefficients is now terminated with -1.
9147 Previously it was terminated with 0, which was also part of the
9148 value; thus, the array representation was not applicable to
9149 polynomials where t^0 has coefficient zero. This change makes
9150 the array representation useful in a more general context.
9154 * Various modifications and fixes to SSL/TLS cipher string
9155 handling. For ECC, the code now distinguishes between fixed ECDH
9156 with RSA certificates on the one hand and with ECDSA certificates
9157 on the other hand, since these are separate ciphersuites. The
9158 unused code for Fortezza ciphersuites has been removed.
9160 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9161 (not "ECDHE"). For consistency with the code for DH
9162 certificates, use of ECDH certificates is now considered ECDH
9163 authentication, not RSA or ECDSA authentication (the latter is
9164 merely the CA's signing algorithm and not actively used in the
9167 The temporary ciphersuite alias "ECCdraft" is no longer
9168 available, and ECC ciphersuites are no longer excluded from "ALL"
9169 and "DEFAULT". The following aliases now exist for RFC 4492
9170 ciphersuites, most of these by analogy with the DH case:
9172 kECDHr - ECDH cert, signed with RSA
9173 kECDHe - ECDH cert, signed with ECDSA
9174 kECDH - ECDH cert (signed with either RSA or ECDSA)
9175 kEECDH - ephemeral ECDH
9176 ECDH - ECDH cert or ephemeral ECDH
9182 AECDH - anonymous ECDH
9183 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9187 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9188 Use correct micalg parameters depending on digest(s) in signed message.
9192 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9193 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9197 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9198 an engine to register a method. Add ENGINE lookups for methods and
9199 functional reference processing.
9203 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9204 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9209 * New -resign option to smime utility. This adds one or more signers
9210 to an existing PKCS#7 signedData structure. Also -md option to use an
9211 alternative message digest algorithm for signing.
9215 * Tidy up PKCS#7 routines and add new functions to make it easier to
9216 create PKCS7 structures containing multiple signers. Update smime
9217 application to support multiple signers.
9221 * New -macalg option to pkcs12 utility to allow setting of an alternative
9226 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9227 Reorganize PBE internals to lookup from a static table using NIDs,
9228 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9229 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9230 PRF which will be automatically used with PBES2.
9234 * Replace the algorithm specific calls to generate keys in "req" with the
9239 * Update PKCS#7 enveloped data routines to use new API. This is now
9240 supported by any public key method supporting the encrypt operation. A
9241 ctrl is added to allow the public key algorithm to examine or modify
9242 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9247 * Add a ctrl to asn1 method to allow a public key algorithm to express
9248 a default digest type to use. In most cases this will be SHA1 but some
9249 algorithms (such as GOST) need to specify an alternative digest. The
9250 return value indicates how strong the preference is 1 means optional and
9251 2 is mandatory (that is it is the only supported type). Modify
9252 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9253 use the default md. Update openssl utilities to use the default digest
9254 type for signing if it is not explicitly indicated.
9258 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9259 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9260 signing method from the key type. This effectively removes the link
9261 between digests and public key types.
9265 * Add an OID cross reference table and utility functions. Its purpose is to
9266 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9267 rsaEncryption. This will allow some of the algorithm specific hackery
9268 needed to use the correct OID to be removed.
9272 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9273 structures for PKCS7_sign(). They are now set up by the relevant public
9278 * Add provisional EC pkey method with support for ECDSA and ECDH.
9282 * Add support for key derivation (agreement) in the API, DH method and
9287 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9288 public and private key formats. As a side effect these add additional
9289 command line functionality not previously available: DSA signatures can be
9290 generated and verified using pkeyutl and DH key support and generation in
9297 *Oliver Tappe <zooey@hirschkaefer.de>*
9299 * New make target "install_html_docs" installs HTML renditions of the
9302 *Oliver Tappe <zooey@hirschkaefer.de>*
9304 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9305 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9306 support key and parameter generation and add initial key generation
9307 functionality for RSA.
9311 * Add functions for main EVP_PKEY_method operations. The undocumented
9312 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9313 `EVP_PKEY_{encrypt,decrypt}_old`.
9317 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9318 key API, doesn't do much yet.
9322 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9323 public key algorithms. New option to openssl utility:
9324 "list-public-key-algorithms" to print out info.
9328 * Implement the Supported Elliptic Curves Extension for
9329 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9333 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9334 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9338 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9339 utilities such as rsa, dsa, dsaparam etc except they process any key
9344 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9345 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9346 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9351 * Initial support for pluggable public key ASN1.
9352 De-spaghettify the public key ASN1 handling. Move public and private
9353 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9354 algorithm specific handling to a single module within the relevant
9355 algorithm directory. Add functions to allow (near) opaque processing
9356 of public and private key structures.
9360 * Implement the Supported Point Formats Extension for
9361 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9365 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9366 for the psk identity [hint] and the psk callback functions to the
9367 SSL_SESSION, SSL and SSL_CTX structure.
9370 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9374 SSL_CTX_use_psk_identity_hint
9375 SSL_get_psk_identity_hint
9376 SSL_get_psk_identity
9377 SSL_use_psk_identity_hint
9379 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9381 * Add RFC 3161 compliant time stamp request creation, response generation
9382 and response verification functionality.
9384 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9386 * Add initial support for TLS extensions, specifically for the server_name
9387 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9388 have new members for a host name. The SSL data structure has an
9389 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9390 stored in that context to allow for session resumption, even after the
9391 SSL has been switched to a new SSL_CTX in reaction to a client's
9392 server_name extension.
9394 New functions (subject to change):
9396 SSL_get_servername()
9397 SSL_get_servername_type()
9400 New CTRL codes and macros (subject to change):
9402 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9403 - SSL_CTX_set_tlsext_servername_callback()
9404 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9405 - SSL_CTX_set_tlsext_servername_arg()
9406 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9408 openssl s_client has a new '-servername ...' option.
9410 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9411 '-key2 ...', '-servername_fatal' (subject to change). This allows
9412 testing the HostName extension for a specific single host name ('-cert'
9413 and '-key' remain fallbacks for handshakes without HostName
9414 negotiation). If the unrecognized_name alert has to be sent, this by
9415 default is a warning; it becomes fatal with the '-servername_fatal'
9418 *Peter Sylvester, Remy Allais, Christophe Renou*
9420 * Whirlpool hash implementation is added.
9424 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9425 bn(64,32). Because of instruction set limitations it doesn't have
9426 any negative impact on performance. This was done mostly in order
9427 to make it possible to share assembler modules, such as bn_mul_mont
9428 implementations, between 32- and 64-bit builds without hassle.
9432 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9433 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9438 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9439 dedicated Montgomery multiplication procedure, is introduced.
9440 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9441 "64-bit" performance on certain 32-bit targets.
9445 * New option SSL_OP_NO_COMP to disable use of compression selectively
9446 in SSL structures. New SSL ctrl to set maximum send fragment size.
9447 Save memory by setting the I/O buffer sizes dynamically instead of
9448 using the maximum available value.
9452 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9453 in addition to the text details.
9457 * Very, very preliminary EXPERIMENTAL support for printing of general
9458 ASN1 structures. This currently produces rather ugly output and doesn't
9459 handle several customised structures at all.
9463 * Integrated support for PVK file format and some related formats such
9464 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9465 these in the 'rsa' and 'dsa' utilities.
9469 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9473 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9474 place for the (very old) "NETSCAPE" format certificates which are now
9475 handled using new ASN1 code equivalents.
9479 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9480 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9481 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9485 * Modify CRL distribution points extension code to print out previously
9486 unsupported fields. Enhance extension setting code to allow setting of
9491 * Add print and set support for Issuing Distribution Point CRL extension.
9495 * Change 'Configure' script to enable Camellia by default.
9502 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9504 * When rejecting SSL/TLS records due to an incorrect version number, never
9505 update s->server with a new major version number. As of
9506 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9507 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9508 the previous behavior could result in a read attempt at NULL when
9509 receiving specific incorrect SSL/TLS records once record payload
9510 protection is active. ([CVE-2010-0740])
9512 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9514 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9515 could be crashed if the relevant tables were not present (e.g. chrooted).
9517 *Tomas Hoger <thoger@redhat.com>*
9519 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9521 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9523 *Martin Olsson, Neel Mehta*
9525 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9526 accommodate for stack sorting, always a write lock!).
9530 * On some versions of WIN32 Heap32Next is very slow. This can cause
9531 excessive delays in the RAND_poll(): over a minute. As a workaround
9532 include a time check in the inner Heap32Next loop too.
9536 * The code that handled flushing of data in SSL/TLS originally used the
9537 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9538 the problem outlined in PR#1949. The fix suggested there however can
9539 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9540 of Apache). So instead simplify the code to flush unconditionally.
9541 This should be fine since flushing with no data to flush is a no op.
9545 * Handle TLS versions 2.0 and later properly and correctly use the
9546 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9547 off ancient servers have a habit of sticking around for a while...
9551 * Modify compression code so it frees up structures without using the
9552 ex_data callbacks. This works around a problem where some applications
9553 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9554 restarting) then use compression (e.g. SSL with compression) later.
9555 This results in significant per-connection memory leaks and
9556 has caused some security issues including CVE-2008-1678 and
9561 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9562 change when encrypting or decrypting.
9566 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9567 connect and renegotiate with servers which do not support RI.
9568 Until RI is more widely deployed this option is enabled by default.
9572 * Add "missing" ssl ctrls to clear options and mode.
9576 * If client attempts to renegotiate and doesn't support RI respond with
9577 a no_renegotiation alert as required by RFC5746. Some renegotiating
9578 TLS clients will continue a connection gracefully when they receive
9579 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9580 waiting for a server hello which it will never receive. Now we treat a
9581 received no_renegotiation alert as a fatal error. This is because
9582 applications requesting a renegotiation might well expect it to succeed
9583 and would have no code in place to handle the server denying it so the
9584 only safe thing to do is to terminate the connection.
9588 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9589 peer supports secure renegotiation and 0 otherwise. Print out peer
9590 renegotiation support in s_client/s_server.
9594 * Replace the highly broken and deprecated SPKAC certification method with
9595 the updated NID creation version. This should correctly handle UTF8.
9599 * Implement RFC5746. Re-enable renegotiation but require the extension
9600 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9601 turns out to be a bad idea. It has been replaced by
9602 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9603 SSL_CTX_set_options(). This is really not recommended unless you
9604 know what you are doing.
9606 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9608 * Fixes to stateless session resumption handling. Use initial_ctx when
9609 issuing and attempting to decrypt tickets in case it has changed during
9610 servername handling. Use a non-zero length session ID when attempting
9611 stateless session resumption: this makes it possible to determine if
9612 a resumption has occurred immediately after receiving server hello
9613 (several places in OpenSSL subtly assume this) instead of later in
9618 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9619 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9620 fixes for a few places where the return code is not checked
9623 *Julia Lawall <julia@diku.dk>*
9625 * Add --strict-warnings option to Configure script to include devteam
9626 warnings in other configurations.
9630 * Add support for --libdir option and LIBDIR variable in makefiles. This
9631 makes it possible to install openssl libraries in locations which
9632 have names other than "lib", for example "/usr/lib64" which some
9635 *Steve Henson, based on patch from Jeremy Utley*
9637 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9638 X690 8.9.12 and can produce some misleading textual output of OIDs.
9640 *Steve Henson, reported by Dan Kaminsky*
9642 * Delete MD2 from algorithm tables. This follows the recommendation in
9643 several standards that it is not used in new applications due to
9644 several cryptographic weaknesses. For binary compatibility reasons
9645 the MD2 API is still compiled in by default.
9649 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9654 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9655 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9658 *Guenter <lists@gknw.net>*
9660 * Fix the server certificate chain building code to use X509_verify_cert(),
9661 it used to have an ad-hoc builder which was unable to cope with anything
9662 other than a simple chain.
9664 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9666 * Don't check self signed certificate signatures in X509_verify_cert()
9667 by default (a flag can override this): it just wastes time without
9668 adding any security. As a useful side effect self signed root CAs
9669 with non-FIPS digests are now usable in FIPS mode.
9673 * In dtls1_process_out_of_seq_message() the check if the current message
9674 is already buffered was missing. For every new message was memory
9675 allocated, allowing an attacker to perform an denial of service attack
9676 with sending out of seq handshake messages until there is no memory
9677 left. Additionally every future message was buffered, even if the
9678 sequence number made no sense and would be part of another handshake.
9679 So only messages with sequence numbers less than 10 in advance will be
9680 buffered. ([CVE-2009-1378])
9682 *Robin Seggelmann, discovered by Daniel Mentz*
9684 * Records are buffered if they arrive with a future epoch to be
9685 processed after finishing the corresponding handshake. There is
9686 currently no limitation to this buffer allowing an attacker to perform
9687 a DOS attack with sending records with future epochs until there is no
9688 memory left. This patch adds the pqueue_size() function to determine
9689 the size of a buffer and limits the record buffer to 100 entries.
9692 *Robin Seggelmann, discovered by Daniel Mentz*
9694 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9695 parent structure is freed. ([CVE-2009-1379])
9699 * Handle non-blocking I/O properly in SSL_shutdown() call.
9701 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9703 * Add `2.5.4.*` OIDs
9705 *Ilya O. <vrghost@gmail.com>*
9707 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9709 * Disable renegotiation completely - this fixes a severe security
9710 problem ([CVE-2009-3555]) at the cost of breaking all
9711 renegotiation. Renegotiation can be re-enabled by setting
9712 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9713 run-time. This is really not recommended unless you know what
9718 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9720 * Don't set val to NULL when freeing up structures, it is freed up by
9721 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9722 zeroing past the valid field. ([CVE-2009-0789])
9724 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9726 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9727 checked correctly. This would allow some invalid signed attributes to
9728 appear to verify correctly. ([CVE-2009-0591])
9730 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9732 * Reject UniversalString and BMPString types with invalid lengths. This
9733 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9734 a legal length. ([CVE-2009-0590])
9738 * Set S/MIME signing as the default purpose rather than setting it
9739 unconditionally. This allows applications to override it at the store
9744 * Permit restricted recursion of ASN1 strings. This is needed in practice
9745 to handle some structures.
9749 * Improve efficiency of mem_gets: don't search whole buffer each time
9752 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9754 * New -hex option for openssl rand.
9758 * Print out UTF8String and NumericString when parsing ASN1.
9762 * Support NumericString type for name components.
9766 * Allow CC in the environment to override the automatically chosen
9767 compiler. Note that nothing is done to ensure flags work with the
9772 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9774 * Properly check EVP_VerifyFinal() and similar return values
9777 *Ben Laurie, Bodo Moeller, Google Security Team*
9779 * Enable TLS extensions by default.
9783 * Allow the CHIL engine to be loaded, whether the application is
9784 multithreaded or not. (This does not release the developer from the
9785 obligation to set up the dynamic locking callbacks.)
9787 *Sander Temme <sander@temme.net>*
9789 * Use correct exit code if there is an error in dgst command.
9791 *Steve Henson; problem pointed out by Roland Dirlewanger*
9793 * Tweak Configure so that you need to say "experimental-jpake" to enable
9794 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9798 * Add experimental JPAKE support, including demo authentication in
9799 s_client and s_server.
9803 * Set the comparison function in v3_addr_canonize().
9805 *Rob Austein <sra@hactrn.net>*
9807 * Add support for XMPP STARTTLS in s_client.
9809 *Philip Paeps <philip@freebsd.org>*
9811 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9812 to ensure that even with this option, only ciphersuites in the
9813 server's preference list will be accepted. (Note that the option
9814 applies only when resuming a session, so the earlier behavior was
9815 just about the algorithm choice for symmetric cryptography.)
9819 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9821 * Fix NULL pointer dereference if a DTLS server received
9822 ChangeCipherSpec as first record ([CVE-2009-1386]).
9826 * Fix a state transition in s3_srvr.c and d1_srvr.c
9827 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9831 * The fix in 0.9.8c that supposedly got rid of unsafe
9832 double-checked locking was incomplete for RSA blinding,
9833 addressing just one layer of what turns out to have been
9834 doubly unsafe triple-checked locking.
9836 So now fix this for real by retiring the MONT_HELPER macro
9837 in crypto/rsa/rsa_eay.c.
9839 *Bodo Moeller; problem pointed out by Marius Schilder*
9841 * Various precautionary measures:
9843 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9845 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9846 (NB: This would require knowledge of the secret session ticket key
9847 to exploit, in which case you'd be SOL either way.)
9849 - Change bn_nist.c so that it will properly handle input BIGNUMs
9850 outside the expected range.
9852 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9855 *Neel Mehta, Bodo Moeller*
9857 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9858 the load fails. Useful for distros.
9860 *Ben Laurie and the FreeBSD team*
9862 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9866 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9870 * Expand ENGINE to support engine supplied SSL client certificate functions.
9872 This work was sponsored by Logica.
9876 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9877 keystores. Support for SSL/TLS client authentication too.
9878 Not compiled unless enable-capieng specified to Configure.
9880 This work was sponsored by Logica.
9884 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9885 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9886 attribute creation routines such as certificate requests and PKCS#12
9891 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9893 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9894 handshake which could lead to a client crash as found using the
9895 Codenomicon TLS test suite ([CVE-2008-1672])
9897 *Steve Henson, Mark Cox*
9899 * Fix double free in TLS server name extensions which could lead to
9900 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9904 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9906 Clear the error queue to ensure that error entries left from
9907 older function calls do not interfere with the correct operation.
9909 *Lutz Jaenicke, Erik de Castro Lopo*
9911 * Remove root CA certificates of commercial CAs:
9913 The OpenSSL project does not recommend any specific CA and does not
9914 have any policy with respect to including or excluding any CA.
9915 Therefore it does not make any sense to ship an arbitrary selection
9916 of root CA certificates with the OpenSSL software.
9920 * RSA OAEP patches to fix two separate invalid memory reads.
9921 The first one involves inputs when 'lzero' is greater than
9922 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9923 before the beginning of from). The second one involves inputs where
9924 the 'db' section contains nothing but zeroes (there is a one-byte
9925 invalid read after the end of 'db').
9927 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9929 * Partial backport from 0.9.9-dev:
9931 Introduce bn_mul_mont (dedicated Montgomery multiplication
9932 procedure) as a candidate for BIGNUM assembler implementation.
9933 While 0.9.9-dev uses assembler for various architectures, only
9934 x86_64 is available by default here in the 0.9.8 branch, and
9935 32-bit x86 is available through a compile-time setting.
9937 To try the 32-bit x86 assembler implementation, use Configure
9938 option "enable-montasm" (which exists only for this backport).
9940 As "enable-montasm" for 32-bit x86 disclaims code stability
9941 anyway, in this constellation we activate additional code
9942 backported from 0.9.9-dev for further performance improvements,
9943 namely BN_from_montgomery_word. (To enable this otherwise,
9944 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9946 *Andy Polyakov (backport partially by Bodo Moeller)*
9948 * Add TLS session ticket callback. This allows an application to set
9949 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9950 values. This is useful for key rollover for example where several key
9951 sets may exist with different names.
9955 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9956 This was broken until now in 0.9.8 releases, such that the only way
9957 a registered ENGINE could be used (assuming it initialises
9958 successfully on the host) was to explicitly set it as the default
9959 for the relevant algorithms. This is in contradiction with 0.9.7
9960 behaviour and the documentation. With this fix, when an ENGINE is
9961 registered into a given algorithm's table of implementations, the
9962 'uptodate' flag is reset so that auto-discovery will be used next
9963 time a new context for that algorithm attempts to select an
9966 *Ian Lister (tweaked by Geoff Thorpe)*
9968 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9969 implementation in the following ways:
9971 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9974 Lack of BER streaming support means one pass streaming processing is
9975 only supported if data is detached: setting the streaming flag is
9976 ignored for embedded content.
9978 CMS support is disabled by default and must be explicitly enabled
9979 with the enable-cms configuration option.
9983 * Update the GMP engine glue to do direct copies between BIGNUM and
9984 mpz_t when openssl and GMP use the same limb size. Otherwise the
9985 existing "conversion via a text string export" trick is still used.
9987 *Paul Sheer <paulsheer@gmail.com>*
9989 * Zlib compression BIO. This is a filter BIO which compressed and
9990 uncompresses any data passed through it.
9994 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9995 RFC3394 compatible AES key wrapping.
9999 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10000 sets string data without copying. X509_ALGOR_set0() and
10001 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10002 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10003 from an X509_ATTRIBUTE structure optionally checking it occurs only
10004 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10009 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10010 to get the expected BN_FLG_CONSTTIME behavior.
10012 *Bodo Moeller (Google)*
10016 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10017 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10018 - added some more tests to do_tests.pl
10019 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10020 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10021 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10022 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10023 - various changes to netware.pl to enable gcc-cross builds on Win32
10025 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10026 - various changes to fix missing prototype warnings
10027 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10028 - added AES, WHIRLPOOL and CPUID assembler code to build files
10029 - added missing AES assembler make rules to mk1mf.pl
10030 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10032 *Guenter Knauf <eflash@gmx.net>*
10034 * Implement certificate status request TLS extension defined in RFC3546.
10035 A client can set the appropriate parameters and receive the encoded
10036 OCSP response via a callback. A server can query the supplied parameters
10037 and set the encoded OCSP response in the callback. Add simplified examples
10038 to s_client and s_server.
10042 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10044 * Fix various bugs:
10045 + Binary incompatibility of ssl_ctx_st structure
10046 + DTLS interoperation with non-compliant servers
10047 + Don't call get_session_cb() without proposed session
10048 + Fix ia64 assembler code
10050 *Andy Polyakov, Steve Henson*
10052 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10054 * DTLS Handshake overhaul. There were longstanding issues with
10055 OpenSSL DTLS implementation, which were making it impossible for
10056 RFC 4347 compliant client to communicate with OpenSSL server.
10057 Unfortunately just fixing these incompatibilities would "cut off"
10058 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10059 server keeps tolerating non RFC compliant syntax. The opposite is
10060 not true, 0.9.8f client can not communicate with earlier server.
10061 This update even addresses CVE-2007-4995.
10065 * Changes to avoid need for function casts in OpenSSL: some compilers
10066 (gcc 4.2 and later) reject their use.
10067 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10070 * Add RFC4507 support to OpenSSL. This includes the corrections in
10071 RFC4507bis. The encrypted ticket format is an encrypted encoded
10072 SSL_SESSION structure, that way new session features are automatically
10075 If a client application caches session in an SSL_SESSION structure
10076 support is transparent because tickets are now stored in the encoded
10079 The SSL_CTX structure automatically generates keys for ticket
10080 protection in servers so again support should be possible
10081 with no application modification.
10083 If a client or server wishes to disable RFC4507 support then the option
10084 SSL_OP_NO_TICKET can be set.
10086 Add a TLS extension debugging callback to allow the contents of any client
10087 or server extensions to be examined.
10089 This work was sponsored by Google.
10093 * Add initial support for TLS extensions, specifically for the server_name
10094 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10095 have new members for a host name. The SSL data structure has an
10096 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10097 stored in that context to allow for session resumption, even after the
10098 SSL has been switched to a new SSL_CTX in reaction to a client's
10099 server_name extension.
10101 New functions (subject to change):
10103 SSL_get_servername()
10104 SSL_get_servername_type()
10107 New CTRL codes and macros (subject to change):
10109 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10110 - SSL_CTX_set_tlsext_servername_callback()
10111 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10112 - SSL_CTX_set_tlsext_servername_arg()
10113 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10115 openssl s_client has a new '-servername ...' option.
10117 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10118 '-key2 ...', '-servername_fatal' (subject to change). This allows
10119 testing the HostName extension for a specific single host name ('-cert'
10120 and '-key' remain fallbacks for handshakes without HostName
10121 negotiation). If the unrecognized_name alert has to be sent, this by
10122 default is a warning; it becomes fatal with the '-servername_fatal'
10125 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10127 * Add AES and SSE2 assembly language support to VC++ build.
10131 * Mitigate attack on final subtraction in Montgomery reduction.
10135 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10136 (which previously caused an internal error).
10140 * Squeeze another 10% out of IGE mode when in != out.
10144 * AES IGE mode speedup.
10146 *Dean Gaudet (Google)*
10148 * Add the Korean symmetric 128-bit cipher SEED (see
10149 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10150 add SEED ciphersuites from RFC 4162:
10152 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10153 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10154 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10155 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10157 To minimize changes between patchlevels in the OpenSSL 0.9.8
10158 series, SEED remains excluded from compilation unless OpenSSL
10159 is configured with 'enable-seed'.
10161 *KISA, Bodo Moeller*
10163 * Mitigate branch prediction attacks, which can be practical if a
10164 single processor is shared, allowing a spy process to extract
10165 information. For detailed background information, see
10166 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10167 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10168 and Necessary Software Countermeasures"). The core of the change
10169 are new versions BN_div_no_branch() and
10170 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10171 respectively, which are slower, but avoid the security-relevant
10172 conditional branches. These are automatically called by BN_div()
10173 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10174 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10175 remove a conditional branch.
10177 BN_FLG_CONSTTIME is the new name for the previous
10178 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10179 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10180 in the exponent causes BN_mod_exp_mont() to use the alternative
10181 implementation in BN_mod_exp_mont_consttime().) The old name
10182 remains as a deprecated alias.
10184 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10185 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10186 constant-time implementations for more than just exponentiation.
10187 Here too the old name is kept as a deprecated alias.
10189 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10190 the BN_BLINDING structure gets an independent copy of the
10191 modulus. This means that the previous `BIGNUM *m` argument to
10192 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10193 essentially becomes `const BIGNUM *m`, although we can't actually
10194 change this in the header file before 0.9.9. It allows
10195 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10196 enable BN_FLG_CONSTTIME.
10198 *Matthew D Wood (Intel Corp)*
10200 * In the SSL/TLS server implementation, be strict about session ID
10201 context matching (which matters if an application uses a single
10202 external cache for different purposes). Previously,
10203 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10204 set. This did ensure strict client verification, but meant that,
10205 with applications using a single external cache for quite
10206 different requirements, clients could circumvent ciphersuite
10207 restrictions for a given session ID context by starting a session
10208 in a different context.
10212 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10213 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10214 authentication-only ciphersuites.
10218 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10219 not complete and could lead to a possible single byte overflow
10220 ([CVE-2007-5135]) [Ben Laurie]
10222 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10224 * Since AES128 and AES256 (and similarly Camellia128 and
10225 Camellia256) share a single mask bit in the logic of
10226 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10227 kludge to work properly if AES128 is available and AES256 isn't
10228 (or if Camellia128 is available and Camellia256 isn't).
10232 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10233 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10234 When a point or a seed is encoded in a BIT STRING, we need to
10235 prevent the removal of trailing zero bits to get the proper DER
10236 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10237 of a NamedBitList, for which trailing 0 bits need to be removed.)
10241 * Have SSL/TLS server implementation tolerate "mismatched" record
10242 protocol version while receiving ClientHello even if the
10243 ClientHello is fragmented. (The server can't insist on the
10244 particular protocol version it has chosen before the ServerHello
10245 message has informed the client about his choice.)
10249 * Add RFC 3779 support.
10251 *Rob Austein for ARIN, Ben Laurie*
10253 * Load error codes if they are not already present instead of using a
10254 static variable. This allows them to be cleanly unloaded and reloaded.
10255 Improve header file function name parsing.
10259 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10260 or CAPABILITY handshake as required by RFCs.
10262 *Goetz Babin-Ebell*
10264 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10266 * Introduce limits to prevent malicious keys being able to
10267 cause a denial of service. ([CVE-2006-2940])
10269 *Steve Henson, Bodo Moeller*
10271 * Fix ASN.1 parsing of certain invalid structures that can result
10272 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10274 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10275 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10277 * Fix SSL client code which could crash if connecting to a
10278 malicious SSLv2 server. ([CVE-2006-4343])
10280 *Tavis Ormandy and Will Drewry, Google Security Team*
10282 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10283 match only those. Before that, "AES256-SHA" would be interpreted
10284 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10285 the same strength classification in 0.9.7h) as we currently only
10286 have a single AES bit in the ciphersuite description bitmap.
10287 That change, however, also applied to ciphersuite strings such as
10288 "RC4-MD5" that intentionally matched multiple ciphersuites --
10289 namely, SSL 2.0 ciphersuites in addition to the more common ones
10290 from SSL 3.0/TLS 1.0.
10292 So we change the selection algorithm again: Naming an explicit
10293 ciphersuite selects this one ciphersuite, and any other similar
10294 ciphersuite (same bitmap) from *other* protocol versions.
10295 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10296 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10298 Since SSL 2.0 does not have any ciphersuites for which the
10299 128/256 bit distinction would be relevant, this works for now.
10300 The proper fix will be to use different bits for AES128 and
10301 AES256, which would have avoided the problems from the beginning;
10302 however, bits are scarce, so we can only do this in a new release
10303 (not just a patchlevel) when we can change the SSL_CIPHER
10304 definition to split the single 'unsigned long mask' bitmap into
10305 multiple values to extend the available space.
10309 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10311 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10312 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10314 * Add AES IGE and biIGE modes.
10318 * Change the Unix randomness entropy gathering to use poll() when
10319 possible instead of select(), since the latter has some
10320 undesirable limitations.
10322 *Darryl Miles via Richard Levitte and Bodo Moeller*
10324 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10325 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10326 cannot be implicitly activated as part of, e.g., the "AES" alias.
10327 However, please upgrade to OpenSSL 0.9.9[-dev] for
10328 non-experimental use of the ECC ciphersuites to get TLS extension
10329 support, which is required for curve and point format negotiation
10330 to avoid potential handshake problems.
10334 * Disable rogue ciphersuites:
10336 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10337 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10338 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10340 The latter two were purportedly from
10341 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10344 Also deactivate the remaining ciphersuites from
10345 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10346 unofficial, and the ID has long expired.
10350 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10351 dual-core machines) and other potential thread-safety issues.
10355 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10356 versions), which is now available for royalty-free use
10357 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10358 Also, add Camellia TLS ciphersuites from RFC 4132.
10360 To minimize changes between patchlevels in the OpenSSL 0.9.8
10361 series, Camellia remains excluded from compilation unless OpenSSL
10362 is configured with 'enable-camellia'.
10366 * Disable the padding bug check when compression is in use. The padding
10367 bug check assumes the first packet is of even length, this is not
10368 necessarily true if compression is enabled and can result in false
10369 positives causing handshake failure. The actual bug test is ancient
10370 code so it is hoped that implementations will either have fixed it by
10371 now or any which still have the bug do not support compression.
10375 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10377 * When applying a cipher rule check to see if string match is an explicit
10378 cipher suite and only match that one cipher suite if it is.
10382 * Link in manifests for VC++ if needed.
10384 *Austin Ziegler <halostatue@gmail.com>*
10386 * Update support for ECC-based TLS ciphersuites according to
10387 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10388 TLS extensions, which are supported starting with the 0.9.9
10389 branch, not in the OpenSSL 0.9.8 branch).
10393 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10394 opaque EVP_CIPHER_CTX handling.
10398 * Fixes and enhancements to zlib compression code. We now only use
10399 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10400 to conform with the standards mentioned here:
10401 <http://www.zlib.net/DLL_FAQ.txt>
10402 Static zlib linking now works on Windows and the new --with-zlib-include
10403 --with-zlib-lib options to Configure can be used to supply the location
10404 of the headers and library. Gracefully handle case where zlib library
10409 * Several fixes and enhancements to the OID generation code. The old code
10410 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10411 handle numbers larger than ULONG_MAX, truncated printing and had a
10412 non standard OBJ_obj2txt() behaviour.
10416 * Add support for building of engines under engine/ as shared libraries
10417 under VC++ build system.
10421 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10422 Hopefully, we will not see any false combination of paths any more.
10426 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10428 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10429 (part of SSL_OP_ALL). This option used to disable the
10430 countermeasure against man-in-the-middle protocol-version
10431 rollback in the SSL 2.0 server implementation, which is a bad
10432 idea. ([CVE-2005-2969])
10434 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10435 for Information Security, National Institute of Advanced Industrial
10436 Science and Technology [AIST], Japan)*
10438 * Add two function to clear and return the verify parameter flags.
10442 * Keep cipherlists sorted in the source instead of sorting them at
10443 runtime, thus removing the need for a lock.
10447 * Avoid some small subgroup attacks in Diffie-Hellman.
10449 *Nick Mathewson and Ben Laurie*
10451 * Add functions for well-known primes.
10455 * Extended Windows CE support.
10457 *Satoshi Nakamura and Andy Polyakov*
10459 * Initialize SSL_METHOD structures at compile time instead of during
10460 runtime, thus removing the need for a lock.
10464 * Make PKCS7_decrypt() work even if no certificate is supplied by
10465 attempting to decrypt each encrypted key in turn. Add support to
10470 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10472 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10475 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10479 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10480 key into the same file any more.
10484 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10488 * Add -utf8 command line and config file option to 'ca'.
10490 *Stefan <stf@udoma.org*
10492 * Removed the macro des_crypt(), as it seems to conflict with some
10493 libraries. Use DES_crypt().
10497 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10498 involves renaming the source and generated shared-libs for
10499 both. The engines will accept the corrected or legacy ids
10500 ('ncipher' and '4758_cca' respectively) when binding. NB,
10501 this only applies when building 'shared'.
10503 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10505 * Add attribute functions to EVP_PKEY structure. Modify
10506 PKCS12_create() to recognize a CSP name attribute and
10507 use it. Make -CSP option work again in pkcs12 utility.
10511 * Add new functionality to the bn blinding code:
10512 - automatic re-creation of the BN_BLINDING parameters after
10513 a fixed number of uses (currently 32)
10514 - add new function for parameter creation
10515 - introduce flags to control the update behaviour of the
10516 BN_BLINDING parameters
10517 - hide BN_BLINDING structure
10518 Add a second BN_BLINDING slot to the RSA structure to improve
10519 performance when a single RSA object is shared among several
10524 * Add support for DTLS.
10526 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10528 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10529 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10533 * Remove buggy and incomplete DH cert support from
10534 ssl/ssl_rsa.c and ssl/s3_both.c
10538 * Use SHA-1 instead of MD5 as the default digest algorithm for
10539 the `apps/openssl` commands.
10543 * Compile clean with "-Wall -Wmissing-prototypes
10544 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10545 DEBUG_SAFESTACK must also be set.
10549 * Change ./Configure so that certain algorithms can be disabled by default.
10550 The new counterpiece to "no-xxx" is "enable-xxx".
10552 The patented RC5 and MDC2 algorithms will now be disabled unless
10553 "enable-rc5" and "enable-mdc2", respectively, are specified.
10555 (IDEA remains enabled despite being patented. This is because IDEA
10556 is frequently required for interoperability, and there is no license
10557 fee for non-commercial use. As before, "no-idea" can be used to
10558 avoid this algorithm.)
10562 * Add processing of proxy certificates (see RFC 3820). This work was
10563 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10564 EGEE (Enabling Grids for E-science in Europe).
10568 * RC4 performance overhaul on modern architectures/implementations, such
10569 as Intel P4, IA-64 and AMD64.
10573 * New utility extract-section.pl. This can be used specify an alternative
10574 section number in a pod file instead of having to treat each file as
10575 a separate case in Makefile. This can be done by adding two lines to the
10578 =for comment openssl_section:XXX
10580 The blank line is mandatory.
10584 * New arguments -certform, -keyform and -pass for s_client and s_server
10585 to allow alternative format key and certificate files and passphrase
10590 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10591 update associated structures and add various utility functions.
10593 Add new policy related verify parameters, include policy checking in
10594 standard verify code. Enhance 'smime' application with extra parameters
10595 to support policy checking and print out.
10599 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10600 Nehemiah processors. These extensions support AES encryption in hardware
10601 as well as RNG (though RNG support is currently disabled).
10603 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10605 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10609 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10611 *Andy Polyakov and a number of other people*
10613 * Improved PowerPC platform support. Most notably BIGNUM assembler
10614 implementation contributed by IBM.
10616 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10618 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10619 exponent rather than 'unsigned long'. There is a corresponding change to
10620 the new 'rsa_keygen' element of the RSA_METHOD structure.
10622 *Jelte Jansen, Geoff Thorpe*
10624 * Functionality for creating the initial serial number file is now
10625 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10627 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10628 number file to 1, which is bound to cause problems. To avoid
10629 the problems while respecting compatibility between different 0.9.7
10630 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10631 CA.pl for serial number initialization. With the new release 0.9.8,
10632 we can fix the problem directly in the 'ca' utility.)
10636 * Reduced header interdependencies by declaring more opaque objects in
10637 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10638 give fewer recursive includes, which could break lazy source code - so
10639 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10640 developers should define this symbol when building and using openssl to
10641 ensure they track the recommended behaviour, interfaces, [etc], but
10642 backwards-compatible behaviour prevails when this isn't defined.
10646 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10650 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10651 This will generate a random key of the appropriate length based on the
10652 cipher context. The EVP_CIPHER can provide its own random key generation
10653 routine to support keys of a specific form. This is used in the des and
10654 3des routines to generate a key of the correct parity. Update S/MIME
10655 code to use new functions and hence generate correct parity DES keys.
10656 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10657 valid (weak or incorrect parity).
10661 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10662 as looking them up. This is useful when the verified structure may contain
10663 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10664 present unless the new PKCS7_NO_CRL flag is asserted.
10668 * Extend ASN1 oid configuration module. It now additionally accepts the
10671 shortName = some long name, 1.2.3.4
10675 * Reimplemented the BN_CTX implementation. There is now no more static
10676 limitation on the number of variables it can handle nor the depth of the
10677 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10678 information can now expand as required, and rather than having a single
10679 static array of bignums, BN_CTX now uses a linked-list of such arrays
10680 allowing it to expand on demand whilst maintaining the usefulness of
10681 BN_CTX's "bundling".
10685 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10686 to allow all RSA operations to function using a single BN_CTX.
10690 * Preliminary support for certificate policy evaluation and checking. This
10691 is initially intended to pass the tests outlined in "Conformance Testing
10692 of Relying Party Client Certificate Path Processing Logic" v1.07.
10696 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10697 remained unused and not that useful. A variety of other little bignum
10698 tweaks and fixes have also been made continuing on from the audit (see
10703 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10704 associated ASN1, EVP and SSL functions and old ASN1 macros.
10708 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10709 and this should never fail. So the return value from the use of
10710 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10711 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10715 * BN_CTX_get() should return zero-valued bignums, providing the same
10716 initialised value as BN_new().
10718 *Geoff Thorpe, suggested by Ulf Möller*
10720 * Support for inhibitAnyPolicy certificate extension.
10724 * An audit of the BIGNUM code is underway, for which debugging code is
10725 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10726 is considered valid when processing BIGNUMs, and causes execution to
10727 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10728 further steps are taken to deliberately pollute unused data in BIGNUM
10729 structures to try and expose faulty code further on. For now, openssl will
10730 (in its default mode of operation) continue to tolerate the inconsistent
10731 forms that it has tolerated in the past, but authors and packagers should
10732 consider trying openssl and their own applications when compiled with
10733 these debugging symbols defined. It will help highlight potential bugs in
10734 their own code, and will improve the test coverage for OpenSSL itself. At
10735 some point, these tighter rules will become openssl's default to improve
10736 maintainability, though the assert()s and other overheads will remain only
10737 in debugging configurations. See bn.h for more details.
10739 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10741 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10742 that can only be obtained through BN_CTX_new() (which implicitly
10743 initialises it). The presence of this function only made it possible
10744 to overwrite an existing structure (and cause memory leaks).
10748 * Because of the callback-based approach for implementing LHASH as a
10749 template type, lh_insert() adds opaque objects to hash-tables and
10750 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10751 to clean up those corresponding objects before destroying the hash table
10752 (and losing the object pointers). So some over-zealous constifications in
10753 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10754 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10755 prototyped to have "const" restrictions on the object pointers they are
10756 given (and so aren't required to cast them away any more).
10760 * The tmdiff.h API was so ugly and minimal that our own timing utility
10761 (speed) prefers to use its own implementation. The two implementations
10762 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10763 its object type properly exposed (MS_TM) instead of casting to/from
10764 `char *`. This may still change yet if someone realises MS_TM and
10766 aren't necessarily the greatest nomenclatures - but this is what was used
10767 internally to the implementation so I've used that for now.
10771 * Ensure that deprecated functions do not get compiled when
10772 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10773 the self-tests were still using deprecated key-generation functions so
10774 these have been updated also.
10778 * Reorganise PKCS#7 code to separate the digest location functionality
10779 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10780 New function PKCS7_set_digest() to set the digest type for PKCS#7
10781 digestedData type. Add additional code to correctly generate the
10782 digestedData type and add support for this type in PKCS7 initialization
10787 * New function PKCS7_set0_type_other() this initializes a PKCS7
10788 structure of type "other".
10792 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10793 sure the loop does correctly stop and breaking ("division by zero")
10794 modulus operations are not performed. The (pre-generated) prime
10795 table crypto/bn/bn_prime.h was already correct, but it could not be
10796 re-generated on some platforms because of the "division by zero"
10797 situation in the script.
10799 *Ralf S. Engelschall*
10801 * Update support for ECC-based TLS ciphersuites according to
10802 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10803 SHA-1 now is only used for "small" curves (where the
10804 representation of a field element takes up to 24 bytes); for
10805 larger curves, the field element resulting from ECDH is directly
10806 used as premaster secret.
10808 *Douglas Stebila (Sun Microsystems Laboratories)*
10810 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10811 curve secp160r1 to the tests.
10813 *Douglas Stebila (Sun Microsystems Laboratories)*
10815 * Add the possibility to load symbols globally with DSO.
10817 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10819 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10820 control of the error stack.
10824 * Add support for STORE in ENGINE.
10828 * Add the STORE type. The intention is to provide a common interface
10829 to certificate and key stores, be they simple file-based stores, or
10830 HSM-type store, or LDAP stores, or...
10831 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10835 * Add a generic structure called OPENSSL_ITEM. This can be used to
10836 pass a list of arguments to any function as well as provide a way
10837 for a function to pass data back to the caller.
10841 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10842 works like BUF_strdup() but can be used to duplicate a portion of
10843 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10848 * Add the function sk_find_ex() which works like sk_find(), but will
10849 return an index to an element even if an exact match couldn't be
10850 found. The index is guaranteed to point at the element where the
10851 searched-for key would be inserted to preserve sorting order.
10855 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10856 takes an extra flags argument for optional functionality. Currently,
10857 the following flags are defined:
10859 OBJ_BSEARCH_VALUE_ON_NOMATCH
10860 This one gets OBJ_bsearch_ex() to return a pointer to the first
10861 element where the comparing function returns a negative or zero
10864 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10865 This one gets OBJ_bsearch_ex() to return a pointer to the first
10866 element where the comparing function returns zero. This is useful
10867 if there are more than one element where the comparing function
10872 * Make it possible to create self-signed certificates with 'openssl ca'
10873 in such a way that the self-signed certificate becomes part of the
10874 CA database and uses the same mechanisms for serial number generation
10875 as all other certificate signing. The new flag '-selfsign' enables
10876 this functionality. Adapt CA.sh and CA.pl.in.
10880 * Add functionality to check the public key of a certificate request
10881 against a given private. This is useful to check that a certificate
10882 request can be signed by that key (self-signing).
10886 * Make it possible to have multiple active certificates with the same
10887 subject in the CA index file. This is done only if the keyword
10888 'unique_subject' is set to 'no' in the main CA section (default
10889 if 'CA_default') of the configuration file. The value is saved
10890 with the database itself in a separate index attribute file,
10891 named like the index file with '.attr' appended to the name.
10895 * Generate multi-valued AVAs using '+' notation in config files for
10900 * Support for nameConstraints certificate extension.
10904 * Support for policyConstraints certificate extension.
10908 * Support for policyMappings certificate extension.
10912 * Make sure the default DSA_METHOD implementation only uses its
10913 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10914 and change its own handlers to be NULL so as to remove unnecessary
10915 indirection. This lets alternative implementations fallback to the
10916 default implementation more easily.
10920 * Support for directoryName in GeneralName related extensions
10925 * Make it possible to link applications using Makefile.shared.
10926 Make that possible even when linking against static libraries!
10930 * Support for single pass processing for S/MIME signing. This now
10931 means that S/MIME signing can be done from a pipe, in addition
10932 cleartext signing (multipart/signed type) is effectively streaming
10933 and the signed data does not need to be all held in memory.
10935 This is done with a new flag PKCS7_STREAM. When this flag is set
10936 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10937 is done after the data is output (and digests calculated) in
10938 SMIME_write_PKCS7().
10942 * Add full support for -rpath/-R, both in shared libraries and
10943 applications, at least on the platforms where it's known how
10948 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10949 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10950 will now compute a table of multiples of the generator that
10951 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10952 faster (notably in the case of a single point multiplication,
10953 scalar * generator).
10955 *Nils Larsch, Bodo Moeller*
10957 * IPv6 support for certificate extensions. The various extensions
10958 which use the IP:a.b.c.d can now take IPv6 addresses using the
10959 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10964 * Added an ENGINE that implements RSA by performing private key
10965 exponentiations with the GMP library. The conversions to and from
10966 GMP's mpz_t format aren't optimised nor are any montgomery forms
10967 cached, and on x86 it appears OpenSSL's own performance has caught up.
10968 However there are likely to be other architectures where GMP could
10969 provide a boost. This ENGINE is not built in by default, but it can be
10970 specified at Configure time and should be accompanied by the necessary
10971 linker additions, eg;
10972 ./config -DOPENSSL_USE_GMP -lgmp
10976 * "openssl engine" will not display ENGINE/DSO load failure errors when
10977 testing availability of engines with "-t" - the old behaviour is
10978 produced by increasing the feature's verbosity with "-tt".
10982 * ECDSA routines: under certain error conditions uninitialized BN objects
10983 could be freed. Solution: make sure initialization is performed early
10984 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10989 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10990 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10991 software implementations. For DSA and DH, parameter generation can
10992 also be overridden by providing the appropriate method callbacks.
10996 * Change the "progress" mechanism used in key-generation and
10997 primality testing to functions that take a new BN_GENCB pointer in
10998 place of callback/argument pairs. The new API functions have `_ex`
10999 postfixes and the older functions are reimplemented as wrappers for
11000 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11001 declarations of the old functions to help (graceful) attempts to
11002 migrate to the new functions. Also, the new key-generation API
11003 functions operate on a caller-supplied key-structure and return
11004 success/failure rather than returning a key or NULL - this is to
11005 help make "keygen" another member function of RSA_METHOD etc.
11007 Example for using the new callback interface:
11009 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11010 void *my_arg = ...;
11013 BN_GENCB_set(&my_cb, my_callback, my_arg);
11015 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11016 /* For the meaning of a, b in calls to my_callback(), see the
11017 * documentation of the function that calls the callback.
11018 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11019 * my_callback should return 1 if it wants BN_is_prime_ex()
11020 * to continue, or 0 to stop.
11025 * Change the ZLIB compression method to be stateful, and make it
11026 available to TLS with the number defined in
11027 draft-ietf-tls-compression-04.txt.
11031 * Add the ASN.1 structures and functions for CertificatePair, which
11032 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11034 CertificatePair ::= SEQUENCE {
11035 forward [0] Certificate OPTIONAL,
11036 reverse [1] Certificate OPTIONAL,
11037 -- at least one of the pair shall be present -- }
11039 Also implement the PEM functions to read and write certificate
11040 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11042 This needed to be defined, mostly for the sake of the LDAP
11043 attribute crossCertificatePair, but may prove useful elsewhere as
11048 * Make it possible to inhibit symlinking of shared libraries in
11049 Makefile.shared, for Cygwin's sake.
11053 * Extend the BIGNUM API by creating a function
11054 void BN_set_negative(BIGNUM *a, int neg);
11055 and a macro that behave like
11056 int BN_is_negative(const BIGNUM *a);
11058 to avoid the need to access 'a->neg' directly in applications.
11062 * Implement fast modular reduction for pseudo-Mersenne primes
11063 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11064 EC_GROUP_new_curve_GFp() will now automatically use this
11067 *Nils Larsch <nla@trustcenter.de>*
11069 * Add new lock type (CRYPTO_LOCK_BN).
11073 * Change the ENGINE framework to automatically load engines
11074 dynamically from specific directories unless they could be
11075 found to already be built in or loaded. Move all the
11076 current engines except for the cryptodev one to a new
11077 directory engines/.
11078 The engines in engines/ are built as shared libraries if
11079 the "shared" options was given to ./Configure or ./config.
11080 Otherwise, they are inserted in libcrypto.a.
11081 /usr/local/ssl/engines is the default directory for dynamic
11082 engines, but that can be overridden at configure time through
11083 the usual use of --prefix and/or --openssldir, and at run
11084 time with the environment variable OPENSSL_ENGINES.
11086 *Geoff Thorpe and Richard Levitte*
11088 * Add Makefile.shared, a helper makefile to build shared
11089 libraries. Adapt Makefile.org.
11093 * Add version info to Win32 DLLs.
11095 *Peter 'Luna' Runestig" <peter@runestig.com>*
11097 * Add new 'medium level' PKCS#12 API. Certificates and keys
11098 can be added using this API to created arbitrary PKCS#12
11099 files while avoiding the low-level API.
11101 New options to PKCS12_create(), key or cert can be NULL and
11102 will then be omitted from the output file. The encryption
11103 algorithm NIDs can be set to -1 for no encryption, the mac
11104 iteration count can be set to 0 to omit the mac.
11106 Enhance pkcs12 utility by making the -nokeys and -nocerts
11107 options work when creating a PKCS#12 file. New option -nomac
11108 to omit the mac, NONE can be set for an encryption algorithm.
11109 New code is modified to use the enhanced PKCS12_create()
11110 instead of the low-level API.
11114 * Extend ASN1 encoder to support indefinite length constructed
11115 encoding. This can output sequences tags and octet strings in
11116 this form. Modify pk7_asn1.c to support indefinite length
11117 encoding. This is experimental and needs additional code to
11118 be useful, such as an ASN1 bio and some enhanced streaming
11121 Extend template encode functionality so that tagging is passed
11122 down to the template encoder.
11126 * Let 'openssl req' fail if an argument to '-newkey' is not
11127 recognized instead of using RSA as a default.
11131 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11132 As these are not official, they are not included in "ALL";
11133 the "ECCdraft" ciphersuite group alias can be used to select them.
11135 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11137 * Add ECDH engine support.
11139 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11141 * Add ECDH in new directory crypto/ecdh/.
11143 *Douglas Stebila (Sun Microsystems Laboratories)*
11145 * Let BN_rand_range() abort with an error after 100 iterations
11146 without success (which indicates a broken PRNG).
11150 * Change BN_mod_sqrt() so that it verifies that the input value
11151 is really the square of the return value. (Previously,
11152 BN_mod_sqrt would show GIGO behaviour.)
11156 * Add named elliptic curves over binary fields from X9.62, SECG,
11157 and WAP/WTLS; add OIDs that were still missing.
11159 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11161 * Extend the EC library for elliptic curves over binary fields
11162 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11165 EC_GF2m_simple_method
11169 EC_GROUP_new_curve_GF2m
11170 EC_GROUP_set_curve_GF2m
11171 EC_GROUP_get_curve_GF2m
11172 EC_POINT_set_affine_coordinates_GF2m
11173 EC_POINT_get_affine_coordinates_GF2m
11174 EC_POINT_set_compressed_coordinates_GF2m
11176 Point compression for binary fields is disabled by default for
11177 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11180 As binary polynomials are represented as BIGNUMs, various members
11181 of the EC_GROUP and EC_POINT data structures can be shared
11182 between the implementations for prime fields and binary fields;
11183 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11184 are essentially identical to their `..._GFp` counterparts.
11185 (For simplicity, the `..._GFp` prefix has been dropped from
11186 various internal method names.)
11188 An internal 'field_div' method (similar to 'field_mul' and
11189 'field_sqr') has been added; this is used only for binary fields.
11191 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11193 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11194 through methods ('mul', 'precompute_mult').
11196 The generic implementations (now internally called 'ec_wNAF_mul'
11197 and 'ec_wNAF_precomputed_mult') remain the default if these
11198 methods are undefined.
11200 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11202 * New function EC_GROUP_get_degree, which is defined through
11203 EC_METHOD. For curves over prime fields, this returns the bit
11204 length of the modulus.
11206 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11208 * New functions EC_GROUP_dup, EC_POINT_dup.
11209 (These simply call ..._new and ..._copy).
11211 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11213 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11214 Polynomials are represented as BIGNUMs (where the sign bit is not
11215 used) in the following functions [macros]:
11218 BN_GF2m_sub [= BN_GF2m_add]
11219 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11220 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11221 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11223 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11224 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11225 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11226 BN_GF2m_cmp [= BN_ucmp]
11228 (Note that only the 'mod' functions are actually for fields GF(2^m).
11229 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11231 For some functions, an the irreducible polynomial defining a
11232 field can be given as an 'unsigned int[]' with strictly
11233 decreasing elements giving the indices of those bits that are set;
11234 i.e., p[] represents the polynomial
11235 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11237 p[0] > p[1] > ... > p[k] = 0.
11238 This applies to the following functions:
11241 BN_GF2m_mod_mul_arr
11242 BN_GF2m_mod_sqr_arr
11243 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11244 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11245 BN_GF2m_mod_exp_arr
11246 BN_GF2m_mod_sqrt_arr
11247 BN_GF2m_mod_solve_quad_arr
11251 Conversion can be performed by the following functions:
11256 bntest.c has additional tests for binary polynomial arithmetic.
11258 Two implementations for BN_GF2m_mod_div() are available.
11259 The default algorithm simply uses BN_GF2m_mod_inv() and
11260 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11261 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11262 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11264 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11266 * Add new error code 'ERR_R_DISABLED' that can be used when some
11267 functionality is disabled at compile-time.
11269 *Douglas Stebila <douglas.stebila@sun.com>*
11271 * Change default behaviour of 'openssl asn1parse' so that more
11272 information is visible when viewing, e.g., a certificate:
11274 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11275 mode the content of non-printable OCTET STRINGs is output in a
11276 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11277 avoid the appearance of a printable string.
11279 *Nils Larsch <nla@trustcenter.de>*
11281 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11283 EC_GROUP_set_asn1_flag()
11284 EC_GROUP_get_asn1_flag()
11285 EC_GROUP_set_point_conversion_form()
11286 EC_GROUP_get_point_conversion_form()
11287 These control ASN1 encoding details:
11288 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11289 has been set to OPENSSL_EC_NAMED_CURVE.
11290 - Points are encoded in uncompressed form by default; options for
11291 asn1_for are as for point2oct, namely
11292 POINT_CONVERSION_COMPRESSED
11293 POINT_CONVERSION_UNCOMPRESSED
11294 POINT_CONVERSION_HYBRID
11296 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11298 EC_GROUP_set_seed()
11299 EC_GROUP_get0_seed()
11300 EC_GROUP_get_seed_len()
11301 This is used only for ASN1 purposes (so far).
11303 *Nils Larsch <nla@trustcenter.de>*
11305 * Add 'field_type' member to EC_METHOD, which holds the NID
11306 of the appropriate field type OID. The new function
11307 EC_METHOD_get_field_type() returns this value.
11309 *Nils Larsch <nla@trustcenter.de>*
11312 EC_POINT_point2bn()
11313 EC_POINT_bn2point()
11314 EC_POINT_point2hex()
11315 EC_POINT_hex2point()
11316 providing useful interfaces to EC_POINT_point2oct() and
11317 EC_POINT_oct2point().
11319 *Nils Larsch <nla@trustcenter.de>*
11321 * Change internals of the EC library so that the functions
11322 EC_GROUP_set_generator()
11323 EC_GROUP_get_generator()
11324 EC_GROUP_get_order()
11325 EC_GROUP_get_cofactor()
11326 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11327 to methods, which would lead to unnecessary code duplication when
11328 adding different types of curves.
11330 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11332 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11333 arithmetic, and such that modified wNAFs are generated
11334 (which avoid length expansion in many cases).
11338 * Add a function EC_GROUP_check_discriminant() (defined via
11339 EC_METHOD) that verifies that the curve discriminant is non-zero.
11341 Add a function EC_GROUP_check() that makes some sanity tests
11342 on a EC_GROUP, its generator and order. This includes
11343 EC_GROUP_check_discriminant().
11345 *Nils Larsch <nla@trustcenter.de>*
11347 * Add ECDSA in new directory crypto/ecdsa/.
11349 Add applications 'openssl ecparam' and 'openssl ecdsa'
11350 (these are based on 'openssl dsaparam' and 'openssl dsa').
11352 ECDSA support is also included in various other files across the
11353 library. Most notably,
11354 - 'openssl req' now has a '-newkey ecdsa:file' option;
11355 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11356 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11357 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11358 them suitable for ECDSA where domain parameters must be
11359 extracted before the specific public key;
11360 - ECDSA engine support has been added.
11362 *Nils Larsch <nla@trustcenter.de>*
11364 * Include some named elliptic curves, and add OIDs from X9.62,
11365 SECG, and WAP/WTLS. Each curve can be obtained from the new
11367 EC_GROUP_new_by_curve_name(),
11368 and the list of available named curves can be obtained with
11369 EC_get_builtin_curves().
11370 Also add a 'curve_name' member to EC_GROUP objects, which can be
11372 EC_GROUP_set_curve_name()
11373 EC_GROUP_get_curve_name()
11375 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11377 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11378 was actually never needed) and in BN_mul(). The removal in BN_mul()
11379 required a small change in bn_mul_part_recursive() and the addition
11380 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11381 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11382 bn_sub_words() and bn_add_words() except they take arrays with
11387 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11389 * Cleanse PEM buffers before freeing them since they may contain
11392 *Benjamin Bennett <ben@psc.edu>*
11394 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11395 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11396 authentication-only ciphersuites.
11400 * Since AES128 and AES256 share a single mask bit in the logic of
11401 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11402 kludge to work properly if AES128 is available and AES256 isn't.
11406 * Expand security boundary to match 1.1.1 module.
11410 * Remove redundant features: hash file source, editing of test vectors
11411 modify fipsld to use external fips_premain.c signature.
11415 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11416 run algorithm test programs.
11420 * Make algorithm test programs more tolerant of whitespace.
11424 * Have SSL/TLS server implementation tolerate "mismatched" record
11425 protocol version while receiving ClientHello even if the
11426 ClientHello is fragmented. (The server can't insist on the
11427 particular protocol version it has chosen before the ServerHello
11428 message has informed the client about his choice.)
11432 * Load error codes if they are not already present instead of using a
11433 static variable. This allows them to be cleanly unloaded and reloaded.
11437 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11439 * Introduce limits to prevent malicious keys being able to
11440 cause a denial of service. ([CVE-2006-2940])
11442 *Steve Henson, Bodo Moeller*
11444 * Fix ASN.1 parsing of certain invalid structures that can result
11445 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11447 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11448 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11450 * Fix SSL client code which could crash if connecting to a
11451 malicious SSLv2 server. ([CVE-2006-4343])
11453 *Tavis Ormandy and Will Drewry, Google Security Team*
11455 * Change ciphersuite string processing so that an explicit
11456 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11457 will no longer include "AES128-SHA"), and any other similar
11458 ciphersuite (same bitmap) from *other* protocol versions (so that
11459 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11460 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11461 changes from 0.9.8b and 0.9.8d.
11465 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11467 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11468 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11470 * Change the Unix randomness entropy gathering to use poll() when
11471 possible instead of select(), since the latter has some
11472 undesirable limitations.
11474 *Darryl Miles via Richard Levitte and Bodo Moeller*
11476 * Disable rogue ciphersuites:
11478 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11479 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11480 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11482 The latter two were purportedly from
11483 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11486 Also deactivate the remaining ciphersuites from
11487 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11488 unofficial, and the ID has long expired.
11492 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11493 dual-core machines) and other potential thread-safety issues.
11497 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11499 * Adapt fipsld and the build system to link against the validated FIPS
11500 module in FIPS mode.
11504 * Fixes for VC++ 2005 build under Windows.
11508 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11509 from a Windows bash shell such as MSYS. It is autodetected from the
11510 "config" script when run from a VC++ environment. Modify standard VC++
11511 build to use fipscanister.o from the GNU make build.
11515 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11517 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11518 The value now differs depending on if you build for FIPS or not.
11519 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11520 safely run with a non-FIPSed libcrypto, as it may crash because of
11521 the difference induced by this change.
11525 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11527 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11528 (part of SSL_OP_ALL). This option used to disable the
11529 countermeasure against man-in-the-middle protocol-version
11530 rollback in the SSL 2.0 server implementation, which is a bad
11531 idea. ([CVE-2005-2969])
11533 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11534 for Information Security, National Institute of Advanced Industrial
11535 Science and Technology [AIST, Japan)]*
11537 * Minimal support for X9.31 signatures and PSS padding modes. This is
11538 mainly for FIPS compliance and not fully integrated at this stage.
11542 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11543 the exponentiation using a fixed-length exponent. (Otherwise,
11544 the information leaked through timing could expose the secret key
11545 after many signatures; cf. Bleichenbacher's attack on DSA with
11550 * Make a new fixed-window mod_exp implementation the default for
11551 RSA, DSA, and DH private-key operations so that the sequence of
11552 squares and multiplies and the memory access pattern are
11553 independent of the particular secret key. This will mitigate
11554 cache-timing and potential related attacks.
11556 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11557 and this is automatically used by BN_mod_exp_mont() if the new flag
11558 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11559 will use this BN flag for private exponents unless the flag
11560 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11561 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11563 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11565 * Change the client implementation for SSLv23_method() and
11566 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11567 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11568 (Previously, the SSL 2.0 backwards compatible Client Hello
11569 message format would be used even with SSL_OP_NO_SSLv2.)
11573 * Add support for smime-type MIME parameter in S/MIME messages which some
11578 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11579 a threadsafe manner. Modify rsa code to use new function and add calls
11580 to dsa and dh code (which had race conditions before).
11584 * Include the fixed error library code in the C error file definitions
11585 instead of fixing them up at runtime. This keeps the error code
11586 structures constant.
11590 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11592 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11595 * Fixes for newer kerberos headers. NB: the casts are needed because
11596 the 'length' field is signed on one version and unsigned on another
11597 with no (?) obvious way to tell the difference, without these VC++
11598 complains. Also the "definition" of FAR (blank) is no longer included
11599 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11600 some needed definitions.
11604 * Undo Cygwin change.
11608 * Added support for proxy certificates according to RFC 3820.
11609 Because they may be a security thread to unaware applications,
11610 they must be explicitly allowed in run-time. See
11611 docs/HOWTO/proxy_certificates.txt for further information.
11615 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11617 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11618 server and client random values. Previously
11619 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11620 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11622 This change has negligible security impact because:
11624 1. Server and client random values still have 24 bytes of pseudo random
11627 2. Server and client random values are sent in the clear in the initial
11630 3. The master secret is derived using the premaster secret (48 bytes in
11631 size for static RSA ciphersuites) as well as client server and random
11634 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11637 *Stephen Henson, reported by UK NISCC*
11639 * Use Windows randomness collection on Cygwin.
11643 * Fix hang in EGD/PRNGD query when communication socket is closed
11644 prematurely by EGD/PRNGD.
11646 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11648 * Prompt for pass phrases when appropriate for PKCS12 input format.
11652 * Back-port of selected performance improvements from development
11653 branch, as well as improved support for PowerPC platforms.
11657 * Add lots of checks for memory allocation failure, error codes to indicate
11658 failure and freeing up memory if a failure occurs.
11660 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11662 * Add new -passin argument to dgst.
11666 * Perform some character comparisons of different types in X509_NAME_cmp:
11667 this is needed for some certificates that re-encode DNs into UTF8Strings
11668 (in violation of RFC3280) and can't or won't issue name rollover
11673 * Make an explicit check during certificate validation to see that
11674 the CA setting in each certificate on the chain is correct. As a
11675 side effect always do the following basic checks on extensions,
11676 not just when there's an associated purpose to the check:
11678 - if there is an unhandled critical extension (unless the user
11679 has chosen to ignore this fault)
11680 - if the path length has been exceeded (if one is set at all)
11681 - that certain extensions fit the associated purpose (if one has
11686 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11688 * Avoid a race condition when CRLs are checked in a multi threaded
11689 environment. This would happen due to the reordering of the revoked
11690 entries during signature checking and serial number lookup. Now the
11691 encoding is cached and the serial number sort performed under a lock.
11692 Add new STACK function sk_is_sorted().
11696 * Add Delta CRL to the extension code.
11700 * Various fixes to s3_pkt.c so alerts are sent properly.
11702 *David Holmes <d.holmes@f5.com>*
11704 * Reduce the chances of duplicate issuer name and serial numbers (in
11705 violation of RFC3280) using the OpenSSL certificate creation utilities.
11706 This is done by creating a random 64 bit value for the initial serial
11707 number when a serial number file is created or when a self signed
11708 certificate is created using 'openssl req -x509'. The initial serial
11709 number file is created using 'openssl x509 -next_serial' in CA.pl
11710 rather than being initialized to 1.
11714 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11716 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11717 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11719 *Joe Orton, Steve Henson*
11721 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11724 *Joe Orton, Steve Henson*
11726 * Make it possible to have multiple active certificates with the same
11727 subject in the CA index file. This is done only if the keyword
11728 'unique_subject' is set to 'no' in the main CA section (default
11729 if 'CA_default') of the configuration file. The value is saved
11730 with the database itself in a separate index attribute file,
11731 named like the index file with '.attr' appended to the name.
11735 * X509 verify fixes. Disable broken certificate workarounds when
11736 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11737 keyUsage extension present. Don't accept CRLs with unhandled critical
11738 extensions: since verify currently doesn't process CRL extensions this
11739 rejects a CRL with *any* critical extensions. Add new verify error codes
11744 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11745 A clarification of RFC2560 will require the use of OCTET STRINGs and
11746 some implementations cannot handle the current raw format. Since OpenSSL
11747 copies and compares OCSP nonces as opaque blobs without any attempt at
11748 parsing them this should not create any compatibility issues.
11752 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11753 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11754 this HMAC (and other) operations are several times slower than OpenSSL
11759 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11761 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11763 * Use the correct content when signing type "other".
11767 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11769 * Fix various bugs revealed by running the NISCC test suite:
11771 Stop out of bounds reads in the ASN1 code when presented with
11772 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11774 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11776 If verify callback ignores invalid public key errors don't try to check
11777 certificate signature with the NULL public key.
11781 * New -ignore_err option in ocsp application to stop the server
11782 exiting on the first error in a request.
11786 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11787 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11792 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11793 extra data after the compression methods not only for TLS 1.0
11794 but also for SSL 3.0 (as required by the specification).
11796 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11798 * Change X509_certificate_type() to mark the key as exported/exportable
11799 when it's 512 *bits* long, not 512 bytes.
11803 * Change AES_cbc_encrypt() so it outputs exact multiple of
11804 blocks during encryption.
11808 * Various fixes to base64 BIO and non blocking I/O. On write
11809 flushes were not handled properly if the BIO retried. On read
11810 data was not being buffered properly and had various logic bugs.
11811 This also affects blocking I/O when the data being decoded is a
11816 * Various S/MIME bugfixes and compatibility changes:
11817 output correct application/pkcs7 MIME type if
11818 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11819 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11820 of files as .eml work). Correctly handle very long lines in MIME
11825 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11827 * Countermeasure against the Klima-Pokorny-Rosa extension of
11828 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11829 a protocol version number mismatch like a decryption error
11830 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11834 * Turn on RSA blinding by default in the default implementation
11835 to avoid a timing attack. Applications that don't want it can call
11836 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11837 They would be ill-advised to do so in most cases.
11839 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11841 * Change RSA blinding code so that it works when the PRNG is not
11842 seeded (in this case, the secret RSA exponent is abused as
11843 an unpredictable seed -- if it is not unpredictable, there
11844 is no point in blinding anyway). Make RSA blinding thread-safe
11845 by remembering the creator's thread ID in rsa->blinding and
11846 having all other threads use local one-time blinding factors
11847 (this requires more computation than sharing rsa->blinding, but
11848 avoids excessive locking; and if an RSA object is not shared
11849 between threads, blinding will still be very fast).
11853 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11854 ENGINE as defaults for all supported algorithms irrespective of
11855 the 'flags' parameter. 'flags' is now honoured, so applications
11856 should make sure they are passing it correctly.
11860 * Target "mingw" now allows native Windows code to be generated in
11861 the Cygwin environment as well as with the MinGW compiler.
11865 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11867 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11868 via timing by performing a MAC computation even if incorrect
11869 block cipher padding has been found. This is a countermeasure
11870 against active attacks where the attacker has to distinguish
11871 between bad padding and a MAC verification error. ([CVE-2003-0078])
11873 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11874 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11875 Martin Vuagnoux (EPFL, Ilion)*
11877 * Make the no-err option work as intended. The intention with no-err
11878 is not to have the whole error stack handling routines removed from
11879 libcrypto, it's only intended to remove all the function name and
11880 reason texts, thereby removing some of the footprint that may not
11881 be interesting if those errors aren't displayed anyway.
11883 NOTE: it's still possible for any application or module to have its
11884 own set of error texts inserted. The routines are there, just not
11885 used by default when no-err is given.
11889 * Add support for FreeBSD on IA64.
11891 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11893 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11894 Kerberos function mit_des_cbc_cksum(). Before this change,
11895 the value returned by DES_cbc_cksum() was like the one from
11896 mit_des_cbc_cksum(), except the bytes were swapped.
11898 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11900 * Allow an application to disable the automatic SSL chain building.
11901 Before this a rather primitive chain build was always performed in
11902 ssl3_output_cert_chain(): an application had no way to send the
11903 correct chain if the automatic operation produced an incorrect result.
11905 Now the chain builder is disabled if either:
11907 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11909 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11911 The reasoning behind this is that an application would not want the
11912 auto chain building to take place if extra chain certificates are
11913 present and it might also want a means of sending no additional
11914 certificates (for example the chain has two certificates and the
11919 * Add the possibility to build without the ENGINE framework.
11921 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11923 * Under Win32 gmtime() can return NULL: check return value in
11924 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11928 * DSA routines: under certain error conditions uninitialized BN objects
11929 could be freed. Solution: make sure initialization is performed early
11930 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11931 Nils Larsch <nla@trustcenter.de> via PR#459)
11935 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11936 checked on reconnect on the client side, therefore session resumption
11937 could still fail with a "ssl session id is different" error. This
11938 behaviour is masked when SSL_OP_ALL is used due to
11939 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11940 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11941 followup to PR #377.
11945 * IA-32 assembler support enhancements: unified ELF targets, support
11946 for SCO/Caldera platforms, fix for Cygwin shared build.
11950 * Add support for FreeBSD on sparc64. As a consequence, support for
11951 FreeBSD on non-x86 processors is separate from x86 processors on
11952 the config script, much like the NetBSD support.
11954 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11956 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11958 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11961 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11962 code (06) was taken as the first octet of the session ID and the last
11963 octet was ignored consequently. As a result SSLv2 client side session
11964 caching could not have worked due to the session ID mismatch between
11966 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11971 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11972 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11977 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11978 seems that in spite of existing for more than a year, many application
11979 author have done nothing to provide the necessary callbacks, which
11980 means that this particular engine will not work properly anywhere.
11981 This is a very unfortunate situation which forces us, in the name
11982 of usability, to give the hw_ncipher.c a static lock, which is part
11984 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11985 appear in 0.9.8 or later. We EXPECT application authors to have
11986 dealt properly with this when 0.9.8 is released (unless we actually
11987 make such changes in the libcrypto locking code that changes will
11988 have to be made anyway).
11992 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11993 octets have been read, EOF or an error occurs. Without this change
11994 some truncated ASN1 structures will not produce an error.
11998 * Disable Heimdal support, since it hasn't been fully implemented.
11999 Still give the possibility to force the use of Heimdal, but with
12000 warnings and a request that patches get sent to openssl-dev.
12004 * Add the VC-CE target, introduce the WINCE sysname, and add
12005 INSTALL.WCE and appropriate conditionals to make it build.
12007 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12009 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12010 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12011 edit numbers of the version.
12013 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12015 * Introduce safe string copy and catenation functions
12016 (BUF_strlcpy() and BUF_strlcat()).
12018 *Ben Laurie (CHATS) and Richard Levitte*
12020 * Avoid using fixed-size buffers for one-line DNs.
12022 *Ben Laurie (CHATS)*
12024 * Add BUF_MEM_grow_clean() to avoid information leakage when
12025 resizing buffers containing secrets, and use where appropriate.
12027 *Ben Laurie (CHATS)*
12029 * Avoid using fixed size buffers for configuration file location.
12031 *Ben Laurie (CHATS)*
12033 * Avoid filename truncation for various CA files.
12035 *Ben Laurie (CHATS)*
12037 * Use sizeof in preference to magic numbers.
12039 *Ben Laurie (CHATS)*
12041 * Avoid filename truncation in cert requests.
12043 *Ben Laurie (CHATS)*
12045 * Add assertions to check for (supposedly impossible) buffer
12048 *Ben Laurie (CHATS)*
12050 * Don't cache truncated DNS entries in the local cache (this could
12051 potentially lead to a spoofing attack).
12053 *Ben Laurie (CHATS)*
12055 * Fix various buffers to be large enough for hex/decimal
12056 representations in a platform independent manner.
12058 *Ben Laurie (CHATS)*
12060 * Add CRYPTO_realloc_clean() to avoid information leakage when
12061 resizing buffers containing secrets, and use where appropriate.
12063 *Ben Laurie (CHATS)*
12065 * Add BIO_indent() to avoid much slightly worrying code to do
12068 *Ben Laurie (CHATS)*
12070 * Convert sprintf()/BIO_puts() to BIO_printf().
12072 *Ben Laurie (CHATS)*
12074 * buffer_gets() could terminate with the buffer only half
12077 *Ben Laurie (CHATS)*
12079 * Add assertions to prevent user-supplied crypto functions from
12080 overflowing internal buffers by having large block sizes, etc.
12082 *Ben Laurie (CHATS)*
12084 * New OPENSSL_assert() macro (similar to assert(), but enabled
12087 *Ben Laurie (CHATS)*
12089 * Eliminate unused copy of key in RC4.
12091 *Ben Laurie (CHATS)*
12093 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12095 *Ben Laurie (CHATS)*
12097 * Fix off-by-one error in EGD path.
12099 *Ben Laurie (CHATS)*
12101 * If RANDFILE path is too long, ignore instead of truncating.
12103 *Ben Laurie (CHATS)*
12105 * Eliminate unused and incorrectly sized X.509 structure
12108 *Ben Laurie (CHATS)*
12110 * Eliminate unused and dangerous function knumber().
12112 *Ben Laurie (CHATS)*
12114 * Eliminate unused and dangerous structure, KSSL_ERR.
12116 *Ben Laurie (CHATS)*
12118 * Protect against overlong session ID context length in an encoded
12119 session object. Since these are local, this does not appear to be
12122 *Ben Laurie (CHATS)*
12124 * Change from security patch (see 0.9.6e below) that did not affect
12125 the 0.9.6 release series:
12127 Remote buffer overflow in SSL3 protocol - an attacker could
12128 supply an oversized master key in Kerberos-enabled versions.
12131 *Ben Laurie (CHATS)*
12133 * Change the SSL kerb5 codes to match RFC 2712.
12137 * Make -nameopt work fully for req and add -reqopt switch.
12139 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12141 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12143 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12145 * Make sure tests can be performed even if the corresponding algorithms
12146 have been removed entirely. This was also the last step to make
12147 OpenSSL compilable with DJGPP under all reasonable conditions.
12149 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12151 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12152 to allow version independent disabling of normally unselected ciphers,
12153 which may be activated as a side-effect of selecting a single cipher.
12155 (E.g., cipher list string "RSA" enables ciphersuites that are left
12156 out of "ALL" because they do not provide symmetric encryption.
12157 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12159 *Lutz Jaenicke, Bodo Moeller*
12161 * Add appropriate support for separate platform-dependent build
12162 directories. The recommended way to make a platform-dependent
12163 build directory is the following (tested on Linux), maybe with
12166 # Place yourself outside of the OpenSSL source tree. In
12167 # this example, the environment variable OPENSSL_SOURCE
12168 # is assumed to contain the absolute OpenSSL source directory.
12169 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12170 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12171 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12172 mkdir -p `dirname $F`
12173 ln -s $OPENSSL_SOURCE/$F $F
12176 To be absolutely sure not to disturb the source tree, a "make clean"
12177 is a good thing. If it isn't successful, don't worry about it,
12178 it probably means the source directory is very clean.
12182 * Make sure any ENGINE control commands make local copies of string
12183 pointers passed to them whenever necessary. Otherwise it is possible
12184 the caller may have overwritten (or deallocated) the original string
12185 data when a later ENGINE operation tries to use the stored values.
12187 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12189 * Improve diagnostics in file reading and command-line digests.
12191 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12193 * Add AES modes CFB and OFB to the object database. Correct an
12194 error in AES-CFB decryption.
12198 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12199 allows existing EVP_CIPHER_CTX structures to be reused after
12200 calling `EVP_*Final()`. This behaviour is used by encryption
12201 BIOs and some applications. This has the side effect that
12202 applications must explicitly clean up cipher contexts with
12203 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12207 * Check the values of dna and dnb in bn_mul_recursive before calling
12208 bn_mul_comba (a non zero value means the a or b arrays do not contain
12209 n2 elements) and fallback to bn_mul_normal if either is not zero.
12213 * Fix escaping of non-ASCII characters when using the -subj option
12214 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12218 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12219 form for "surname", serialNumber has no short form.
12220 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12221 therefore remove "mail" short name for "internet 7".
12222 The OID for unique identifiers in X509 certificates is
12223 x500UniqueIdentifier, not uniqueIdentifier.
12224 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12228 * Add an "init" command to the ENGINE config module and auto initialize
12229 ENGINEs. Without any "init" command the ENGINE will be initialized
12230 after all ctrl commands have been executed on it. If init=1 the
12231 ENGINE is initialized at that point (ctrls before that point are run
12232 on the uninitialized ENGINE and after on the initialized one). If
12233 init=0 then the ENGINE will not be initialized at all.
12237 * Fix the 'app_verify_callback' interface so that the user-defined
12238 argument is actually passed to the callback: In the
12239 SSL_CTX_set_cert_verify_callback() prototype, the callback
12240 declaration has been changed from
12243 int (*cb)(X509_STORE_CTX *,void *);
12244 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12245 i=s->ctx->app_verify_callback(&ctx)
12246 has been changed into
12247 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12249 To update applications using SSL_CTX_set_cert_verify_callback(),
12250 a dummy argument can be added to their callback functions.
12252 *D. K. Smetters <smetters@parc.xerox.com>*
12254 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12256 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12258 * Add and OPENSSL_LOAD_CONF define which will cause
12259 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12260 This allows older applications to transparently support certain
12261 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12262 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12263 load the config file and OPENSSL_add_all_algorithms_conf() which will
12264 always load it have also been added.
12268 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12269 Adjust NIDs and EVP layer.
12271 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12273 * Config modules support in openssl utility.
12275 Most commands now load modules from the config file,
12276 though in a few (such as version) this isn't done
12277 because it couldn't be used for anything.
12279 In the case of ca and req the config file used is
12280 the same as the utility itself: that is the -config
12281 command line option can be used to specify an
12286 * Move default behaviour from OPENSSL_config(). If appname is NULL
12287 use "openssl_conf" if filename is NULL use default openssl config file.
12291 * Add an argument to OPENSSL_config() to allow the use of an alternative
12292 config section name. Add a new flag to tolerate a missing config file
12293 and move code to CONF_modules_load_file().
12297 * Support for crypto accelerator cards from Accelerated Encryption
12298 Processing, www.aep.ie. (Use engine 'aep')
12299 The support was copied from 0.9.6c [engine] and adapted/corrected
12300 to work with the new engine framework.
12302 *AEP Inc. and Richard Levitte*
12304 * Support for SureWare crypto accelerator cards from Baltimore
12305 Technologies. (Use engine 'sureware')
12306 The support was copied from 0.9.6c [engine] and adapted
12307 to work with the new engine framework.
12311 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12312 make the newer ENGINE framework commands for the CHIL engine work.
12314 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12316 * Make it possible to produce shared libraries on ReliantUNIX.
12318 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12320 * Add the configuration target debug-linux-ppro.
12321 Make 'openssl rsa' use the general key loading routines
12322 implemented in `apps.c`, and make those routines able to
12323 handle the key format FORMAT_NETSCAPE and the variant
12326 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12328 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12330 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12332 * Add -keyform to rsautl, and document -engine.
12334 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12336 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12337 BIO_R_NO_SUCH_FILE error code rather than the generic
12338 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12342 * Add new functions
12343 ERR_peek_last_error
12344 ERR_peek_last_error_line
12345 ERR_peek_last_error_line_data.
12346 These are similar to
12348 ERR_peek_error_line
12349 ERR_peek_error_line_data,
12350 but report on the latest error recorded rather than the first one
12351 still in the error queue.
12353 *Ben Laurie, Bodo Moeller*
12355 * default_algorithms option in ENGINE config module. This allows things
12357 default_algorithms = ALL
12358 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12362 * Preliminary ENGINE config module.
12366 * New experimental application configuration code.
12370 * Change the AES code to follow the same name structure as all other
12371 symmetric ciphers, and behave the same way. Move everything to
12372 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12374 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12376 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12378 *Ben Laurie and Theo de Raadt*
12380 * Add option to output public keys in req command.
12382 *Massimiliano Pala madwolf@openca.org*
12384 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12385 (up to about 10% better than before for P-192 and P-224).
12389 * New functions/macros
12391 SSL_CTX_set_msg_callback(ctx, cb)
12392 SSL_CTX_set_msg_callback_arg(ctx, arg)
12393 SSL_set_msg_callback(ssl, cb)
12394 SSL_set_msg_callback_arg(ssl, arg)
12396 to request calling a callback function
12398 void cb(int write_p, int version, int content_type,
12399 const void *buf, size_t len, SSL *ssl, void *arg)
12401 whenever a protocol message has been completely received
12402 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12403 protocol version according to which the SSL library interprets
12404 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12405 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12406 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12407 specification (change_cipher_spec(20), alert(21), handshake(22)).
12408 'buf' and 'len' point to the actual message, 'ssl' to the
12409 SSL object, and 'arg' is the application-defined value set by
12410 SSL[_CTX]_set_msg_callback_arg().
12412 'openssl s_client' and 'openssl s_server' have new '-msg' options
12413 to enable a callback that displays all protocol messages.
12417 * Change the shared library support so shared libraries are built as
12418 soon as the corresponding static library is finished, and thereby get
12419 openssl and the test programs linked against the shared library.
12420 This still only happens when the keyword "shard" has been given to
12421 the configuration scripts.
12423 NOTE: shared library support is still an experimental thing, and
12424 backward binary compatibility is still not guaranteed.
12426 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12428 * Add support for Subject Information Access extension.
12430 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12432 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12433 additional bytes when new memory had to be allocated, not just
12434 when reusing an existing buffer.
12438 * New command line and configuration option 'utf8' for the req command.
12439 This allows field values to be specified as UTF8 strings.
12443 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12444 runs for the former and machine-readable output for the latter.
12448 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12449 of the e-mail address in the DN (i.e., it will go into a certificate
12450 extension only). The new configuration file option 'email_in_dn = no'
12451 has the same effect.
12453 *Massimiliano Pala madwolf@openca.org*
12455 * Change all functions with names starting with `des_` to be starting
12456 with `DES_` instead. Add wrappers that are compatible with libdes,
12457 but are named `_ossl_old_des_*`. Finally, add macros that map the
12458 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12459 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12460 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12463 Since we provide two compatibility mappings, the user needs to
12464 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12465 compatibility is desired. The default (i.e., when that macro
12466 isn't defined) is OpenSSL 0.9.6c compatibility.
12468 There are also macros that enable and disable the support of old
12469 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12470 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12471 are defined, the default will apply: to support the old des routines.
12473 In either case, one must include openssl/des.h to get the correct
12474 definitions. Do not try to just include openssl/des_old.h, that
12477 NOTE: This is a major break of an old API into a new one. Software
12478 authors are encouraged to switch to the `DES_` style functions. Some
12479 time in the future, des_old.h and the libdes compatibility functions
12480 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12481 default), and then completely removed.
12485 * Test for certificates which contain unsupported critical extensions.
12486 If such a certificate is found during a verify operation it is
12487 rejected by default: this behaviour can be overridden by either
12488 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12489 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12490 X509_supported_extension() has also been added which returns 1 if a
12491 particular extension is supported.
12495 * Modify the behaviour of EVP cipher functions in similar way to digests
12496 to retain compatibility with existing code.
12500 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12501 compatibility with existing code. In particular the 'ctx' parameter does
12502 not have to be to be initialized before the call to EVP_DigestInit() and
12503 it is tidied up after a call to EVP_DigestFinal(). New function
12504 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12505 EVP_MD_CTX_copy() changed to not require the destination to be
12506 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12507 requires the destination to be valid.
12509 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12510 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12514 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12515 so that complete 'Handshake' protocol structures are kept in memory
12516 instead of overwriting 'msg_type' and 'length' with 'body' data.
12520 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12522 *Massimo Santin via Richard Levitte*
12524 * Major restructuring to the underlying ENGINE code. This includes
12525 reduction of linker bloat, separation of pure "ENGINE" manipulation
12526 (initialisation, etc) from functionality dealing with implementations
12527 of specific crypto interfaces. This change also introduces integrated
12528 support for symmetric ciphers and digest implementations - so ENGINEs
12529 can now accelerate these by providing EVP_CIPHER and EVP_MD
12530 implementations of their own. This is detailed in
12531 [crypto/engine/README.md](crypto/engine/README.md)
12532 as it couldn't be adequately described here. However, there are a few
12533 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12534 were changed in the original introduction of ENGINE code have now
12535 reverted back - the hooking from this code to ENGINE is now a good
12536 deal more passive and at run-time, operations deal directly with
12537 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12538 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12539 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12540 they were not being used by the framework as there is no concept of a
12541 BIGNUM_METHOD and they could not be generalised to the new
12542 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12543 ENGINE_cpy() has been removed as it cannot be consistently defined in
12548 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12552 * Change mkdef.pl to sort symbols that get the same entry number,
12553 and make sure the automatically generated functions `ERR_load_*`
12554 become part of libeay.num as well.
12558 * New function SSL_renegotiate_pending(). This returns true once
12559 renegotiation has been requested (either SSL_renegotiate() call
12560 or HelloRequest/ClientHello received from the peer) and becomes
12561 false once a handshake has been completed.
12562 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12563 sends a HelloRequest, but does not ensure that a handshake takes
12564 place. SSL_renegotiate_pending() is useful for checking if the
12565 client has followed the request.)
12569 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12570 By default, clients may request session resumption even during
12571 renegotiation (if session ID contexts permit); with this option,
12572 session resumption is possible only in the first handshake.
12574 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12575 more bits available for options that should not be part of
12576 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12580 * Add some demos for certificate and certificate request creation.
12584 * Make maximum certificate chain size accepted from the peer application
12585 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12586 "Douglas E. Engert" <deengert@anl.gov>.
12590 * Add support for shared libraries for Unixware-7
12591 (Boyd Lynn Gerber <gerberb@zenez.com>).
12595 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12596 be done prior to destruction. Use this to unload error strings from
12597 ENGINEs that load their own error strings. NB: This adds two new API
12598 functions to "get" and "set" this destroy handler in an ENGINE.
12602 * Alter all existing ENGINE implementations (except "openssl" and
12603 "openbsd") to dynamically instantiate their own error strings. This
12604 makes them more flexible to be built both as statically-linked ENGINEs
12605 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12606 Also, add stub code to each that makes building them as self-contained
12607 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12611 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12612 implementations into applications that are completely implemented in
12613 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12614 commands that can be used to configure what shared-library to load and
12615 to control aspects of the way it is handled. Also, made an update to
12616 the [README-Engine.md](README-Engine.md) file
12617 that brings its information up-to-date and
12618 provides some information and instructions on the "dynamic" ENGINE
12619 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12623 * Make it possible to unload ranges of ERR strings with a new
12624 "ERR_unload_strings" function.
12628 * Add a copy() function to EVP_MD.
12632 * Make EVP_MD routines take a context pointer instead of just the
12633 md_data void pointer.
12637 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12638 that the digest can only process a single chunk of data
12639 (typically because it is provided by a piece of
12640 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12641 is only going to provide a single chunk of data, and hence the
12642 framework needn't accumulate the data for oneshot drivers.
12646 * As with "ERR", make it possible to replace the underlying "ex_data"
12647 functions. This change also alters the storage and management of global
12648 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12649 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12650 index counters. The API functions that use this state have been changed
12651 to take a "class_index" rather than pointers to the class's local STACK
12652 and counter, and there is now an API function to dynamically create new
12653 classes. This centralisation allows us to (a) plug a lot of the
12654 thread-safety problems that existed, and (b) makes it possible to clean
12655 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12656 such data would previously have always leaked in application code and
12657 workarounds were in place to make the memory debugging turn a blind eye
12658 to it. Application code that doesn't use this new function will still
12659 leak as before, but their memory debugging output will announce it now
12660 rather than letting it slide.
12662 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12663 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12664 has a return value to indicate success or failure.
12668 * Make it possible to replace the underlying "ERR" functions such that the
12669 global state (2 LHASH tables and 2 locks) is only used by the "default"
12670 implementation. This change also adds two functions to "get" and "set"
12671 the implementation prior to it being automatically set the first time
12672 any other ERR function takes place. Ie. an application can call "get",
12673 pass the return value to a module it has just loaded, and that module
12674 can call its own "set" function using that value. This means the
12675 module's "ERR" operations will use (and modify) the error state in the
12676 application and not in its own statically linked copy of OpenSSL code.
12680 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12681 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12682 the operation, and provides a more encapsulated way for external code
12683 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12684 to use these functions rather than manually incrementing the counts.
12686 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12690 * Add EVP test program.
12694 * Add symmetric cipher support to ENGINE. Expect the API to change!
12698 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12699 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12700 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12701 These allow a CRL to be built without having to access X509_CRL fields
12702 directly. Modify 'ca' application to use new functions.
12706 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12707 bug workarounds. Rollback attack detection is a security feature.
12708 The problem will only arise on OpenSSL servers when TLSv1 is not
12709 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12710 Software authors not wanting to support TLSv1 will have special reasons
12711 for their choice and can explicitly enable this option.
12713 *Bodo Moeller, Lutz Jaenicke*
12715 * Rationalise EVP so it can be extended: don't include a union of
12716 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12717 (similar to those existing for EVP_CIPHER_CTX).
12722 EVP_MD_CTX_init(&md); /* new function call */
12723 EVP_DigestInit(&md, EVP_sha1());
12724 EVP_DigestUpdate(&md, in, len);
12725 EVP_DigestFinal(&md, out, NULL);
12726 EVP_MD_CTX_cleanup(&md); /* new function call */
12730 * Make DES key schedule conform to the usual scheme, as well as
12731 correcting its structure. This means that calls to DES functions
12732 now have to pass a pointer to a des_key_schedule instead of a
12733 plain des_key_schedule (which was actually always a pointer
12736 des_key_schedule ks;
12738 des_set_key_checked(..., &ks);
12739 des_ncbc_encrypt(..., &ks, ...);
12741 (Note that a later change renames 'des_...' into 'DES_...'.)
12745 * Initial reduction of linker bloat: the use of some functions, such as
12746 PEM causes large amounts of unused functions to be linked in due to
12747 poor organisation. For example pem_all.c contains every PEM function
12748 which has a knock on effect of linking in large amounts of (unused)
12749 ASN1 code. Grouping together similar functions and splitting unrelated
12750 functions prevents this.
12754 * Cleanup of EVP macros.
12758 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12759 correct `_ecb suffix`.
12763 * Add initial OCSP responder support to ocsp application. The
12764 revocation information is handled using the text based index
12765 use by the ca application. The responder can either handle
12766 requests generated internally, supplied in files (for example
12767 via a CGI script) or using an internal minimal server.
12771 * Add configuration choices to get zlib compression for TLS.
12775 * Changes to Kerberos SSL for RFC 2712 compliance:
12776 1. Implemented real KerberosWrapper, instead of just using
12777 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12778 2. Implemented optional authenticator field of KerberosWrapper.
12780 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12781 and authenticator structs; see crypto/krb5/.
12783 Generalized Kerberos calls to support multiple Kerberos libraries.
12784 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12785 via Richard Levitte*
12787 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12788 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12789 values for each of the key sizes rather than having just
12790 parameters (and 'speed' generating keys each time).
12794 * Speed up EVP routines.
12797 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12798 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12799 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12800 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12802 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12803 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12804 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12807 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12809 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12813 * Added the OS2-EMX target.
12815 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12817 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12818 New functions to support `NCONF` routines in extension code.
12819 New function `CONF_set_nconf()`
12820 to allow functions which take an `NCONF` to also handle the old `LHASH`
12821 structure: this means that the old `CONF` compatible routines can be
12822 retained (in particular w.rt. extensions) without having to duplicate the
12823 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12827 * Enhance the general user interface with mechanisms for inner control
12828 and with possibilities to have yes/no kind of prompts.
12832 * Change all calls to low-level digest routines in the library and
12833 applications to use EVP. Add missing calls to HMAC_cleanup() and
12834 don't assume HMAC_CTX can be copied using memcpy().
12836 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12838 * Add the possibility to control engines through control names but with
12839 arbitrary arguments instead of just a string.
12840 Change the key loaders to take a UI_METHOD instead of a callback
12841 function pointer. NOTE: this breaks binary compatibility with earlier
12842 versions of OpenSSL [engine].
12843 Adapt the nCipher code for these new conditions and add a card insertion
12848 * Enhance the general user interface with mechanisms to better support
12849 dialog box interfaces, application-defined prompts, the possibility
12850 to use defaults (for example default passwords from somewhere else)
12851 and interrupts/cancellations.
12855 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12856 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12860 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12861 tidy up some unnecessarily weird code in 'sk_new()').
12863 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12865 * Change the key loading routines for ENGINEs to use the same kind
12866 callback (pem_password_cb) as all other routines that need this
12871 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12872 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12873 than this minimum value is recommended.
12877 * New random seeder for OpenVMS, using the system process statistics
12878 that are easily reachable.
12882 * Windows apparently can't transparently handle global
12883 variables defined in DLLs. Initialisations such as:
12885 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12887 won't compile. This is used by the any applications that need to
12888 declare their own ASN1 modules. This was fixed by adding the option
12889 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12890 needed for static libraries under Win32.
12894 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12895 setting of purpose and trust fields. New X509_STORE trust and
12896 purpose functions and tidy up setting in other SSL functions.
12900 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12901 structure. These are inherited by X509_STORE_CTX when it is
12902 initialised. This allows various defaults to be set in the
12903 X509_STORE structure (such as flags for CRL checking and custom
12904 purpose or trust settings) for functions which only use X509_STORE_CTX
12905 internally such as S/MIME.
12907 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12908 trust settings if they are not set in X509_STORE. This allows X509_STORE
12909 purposes and trust (in S/MIME for example) to override any set by default.
12911 Add command line options for CRL checking to smime, s_client and s_server
12916 * Initial CRL based revocation checking. If the CRL checking flag(s)
12917 are set then the CRL is looked up in the X509_STORE structure and
12918 its validity and signature checked, then if the certificate is found
12919 in the CRL the verify fails with a revoked error.
12921 Various new CRL related callbacks added to X509_STORE_CTX structure.
12923 Command line options added to 'verify' application to support this.
12925 This needs some additional work, such as being able to handle multiple
12926 CRLs with different times, extension based lookup (rather than just
12927 by subject name) and ultimately more complete V2 CRL extension
12932 * Add a general user interface API (crypto/ui/). This is designed
12933 to replace things like des_read_password and friends (backward
12934 compatibility functions using this new API are provided).
12935 The purpose is to remove prompting functions from the DES code
12936 section as well as provide for prompting through dialog boxes in
12937 a window system and the like.
12941 * Add "ex_data" support to ENGINE so implementations can add state at a
12942 per-structure level rather than having to store it globally.
12946 * Make it possible for ENGINE structures to be copied when retrieved by
12947 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12948 This causes the "original" ENGINE structure to act like a template,
12949 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12950 operational state can be localised to each ENGINE structure, despite the
12951 fact they all share the same "methods". New ENGINE structures returned in
12952 this case have no functional references and the return value is the single
12953 structural reference. This matches the single structural reference returned
12954 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12959 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12960 needs to match any other type at all we need to manually clear the
12965 * Changes to the "openssl engine" utility to include;
12966 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12967 about an ENGINE's available control commands.
12968 - executing control commands from command line arguments using the
12969 '-pre' and '-post' switches. '-post' is only used if '-t' is
12970 specified and the ENGINE is successfully initialised. The syntax for
12971 the individual commands are colon-separated, for example;
12972 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12976 * New dynamic control command support for ENGINEs. ENGINEs can now
12977 declare their own commands (numbers), names (strings), descriptions,
12978 and input types for run-time discovery by calling applications. A
12979 subset of these commands are implicitly classed as "executable"
12980 depending on their input type, and only these can be invoked through
12981 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12982 can be based on user input, config files, etc). The distinction is
12983 that "executable" commands cannot return anything other than a boolean
12984 result and can only support numeric or string input, whereas some
12985 discoverable commands may only be for direct use through
12986 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12987 pointers, or other custom uses. The "executable" commands are to
12988 support parameterisations of ENGINE behaviour that can be
12989 unambiguously defined by ENGINEs and used consistently across any
12990 OpenSSL-based application. Commands have been added to all the
12991 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12992 control over shared-library paths without source code alterations.
12996 * Changed all ENGINE implementations to dynamically allocate their
12997 ENGINEs rather than declaring them statically. Apart from this being
12998 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12999 this also allows the implementations to compile without using the
13000 internal engine_int.h header.
13004 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13005 'const' value. Any code that should be able to modify a RAND_METHOD
13006 should already have non-const pointers to it (ie. they should only
13007 modify their own ones).
13011 * Made a variety of little tweaks to the ENGINE code.
13012 - "atalla" and "ubsec" string definitions were moved from header files
13013 to C code. "nuron" string definitions were placed in variables
13014 rather than hard-coded - allowing parameterisation of these values
13015 later on via ctrl() commands.
13016 - Removed unused "#if 0"'d code.
13017 - Fixed engine list iteration code so it uses ENGINE_free() to release
13018 structural references.
13019 - Constified the RAND_METHOD element of ENGINE structures.
13020 - Constified various get/set functions as appropriate and added
13021 missing functions (including a catch-all ENGINE_cpy that duplicates
13022 all ENGINE values onto a new ENGINE except reference counts/state).
13023 - Removed NULL parameter checks in get/set functions. Setting a method
13024 or function to NULL is a way of cancelling out a previously set
13025 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13026 and doesn't justify the extra error symbols and code.
13027 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13028 flags from engine_int.h to engine.h.
13029 - Changed prototypes for ENGINE handler functions (init(), finish(),
13030 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13034 * Implement binary inversion algorithm for BN_mod_inverse in addition
13035 to the algorithm using long division. The binary algorithm can be
13036 used only if the modulus is odd. On 32-bit systems, it is faster
13037 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13038 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13039 up to 450 bits. In 64-bit environments, the binary algorithm
13040 appears to be advantageous for much longer moduli; here we use it
13041 for moduli up to 2048 bits.
13045 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13046 could not support the combine flag in choice fields.
13050 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13051 extensions from a certificate request to the certificate.
13055 * Allow multiple 'certopt' and 'nameopt' options to be separated
13056 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13057 file: this allows the display of the certificate about to be
13058 signed to be customised, to allow certain fields to be included
13059 or excluded and extension details. The old system didn't display
13060 multicharacter strings properly, omitted fields not in the policy
13061 and couldn't display additional details such as extensions.
13065 * Function EC_POINTs_mul for multiple scalar multiplication
13066 of an arbitrary number of elliptic curve points
13067 \sum scalars[i]*points[i],
13068 optionally including the generator defined for the EC_GROUP:
13069 scalar*generator + \sum scalars[i]*points[i].
13071 EC_POINT_mul is a simple wrapper function for the typical case
13072 that the point list has just one item (besides the optional
13077 * First EC_METHODs for curves over GF(p):
13079 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13080 operations and provides various method functions that can also
13081 operate with faster implementations of modular arithmetic.
13083 EC_GFp_mont_method() reuses most functions that are part of
13084 EC_GFp_simple_method, but uses Montgomery arithmetic.
13086 *Bodo Moeller; point addition and point doubling
13087 implementation directly derived from source code provided by
13088 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13090 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13091 crypto/ec/ec_lib.c):
13093 Curves are EC_GROUP objects (with an optional group generator)
13094 based on EC_METHODs that are built into the library.
13096 Points are EC_POINT objects based on EC_GROUP objects.
13098 Most of the framework would be able to handle curves over arbitrary
13099 finite fields, but as there are no obvious types for fields other
13100 than GF(p), some functions are limited to that for now.
13104 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13105 that the file contains a complete HTTP response.
13109 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13110 change the def and num file printf format specifier from "%-40sXXX"
13111 to "%-39s XXX". The latter will always guarantee a space after the
13112 field while the former will cause them to run together if the field
13113 is 40 of more characters long.
13117 * Constify the cipher and digest 'method' functions and structures
13118 and modify related functions to take constant EVP_MD and EVP_CIPHER
13123 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13124 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13128 * Modify `EVP_Digest*()` routines so they now return values. Although the
13129 internal software routines can never fail additional hardware versions
13134 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13136 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13137 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13140 ERR_R_NESTED_ASN1_ERROR
13142 ERR_R_MISSING_ASN1_EOS
13143 were 4 .. 9, conflicting with
13144 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13146 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13147 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13149 Add new error code 'ERR_R_INTERNAL_ERROR'.
13153 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13158 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13159 sets the subject name for a new request or supersedes the
13160 subject name in a given request. Formats that can be parsed are
13161 'CN=Some Name, OU=myOU, C=IT'
13163 'CN=Some Name/OU=myOU/C=IT'.
13165 Add options '-batch' and '-verbose' to 'openssl req'.
13167 *Massimiliano Pala <madwolf@hackmasters.net>*
13169 * Introduce the possibility to access global variables through
13170 functions on platform were that's the best way to handle exporting
13171 global variables in shared libraries. To enable this functionality,
13172 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13173 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13174 is normally done by Configure or something similar).
13176 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13177 in the source file (foo.c) like this:
13179 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13180 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13182 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13183 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13185 OPENSSL_DECLARE_GLOBAL(int,foo);
13186 #define foo OPENSSL_GLOBAL_REF(foo)
13187 OPENSSL_DECLARE_GLOBAL(double,bar);
13188 #define bar OPENSSL_GLOBAL_REF(bar)
13190 The #defines are very important, and therefore so is including the
13191 header file everywhere where the defined globals are used.
13193 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13194 of ASN.1 items, but that structure is a bit different.
13196 The largest change is in util/mkdef.pl which has been enhanced with
13197 better and easier to understand logic to choose which symbols should
13198 go into the Windows .def files as well as a number of fixes and code
13199 cleanup (among others, algorithm keywords are now sorted
13200 lexicographically to avoid constant rewrites).
13204 * In BN_div() keep a copy of the sign of 'num' before writing the
13205 result to 'rm' because if rm==num the value will be overwritten
13206 and produce the wrong result if 'num' is negative: this caused
13207 problems with BN_mod() and BN_nnmod().
13211 * Function OCSP_request_verify(). This checks the signature on an
13212 OCSP request and verifies the signer certificate. The signer
13213 certificate is just checked for a generic purpose and OCSP request
13218 * Add OCSP_check_validity() function to check the validity of OCSP
13219 responses. OCSP responses are prepared in real time and may only
13220 be a few seconds old. Simply checking that the current time lies
13221 between thisUpdate and nextUpdate max reject otherwise valid responses
13222 caused by either OCSP responder or client clock inaccuracy. Instead
13223 we allow thisUpdate and nextUpdate to fall within a certain period of
13224 the current time. The age of the response can also optionally be
13225 checked. Two new options -validity_period and -status_age added to
13230 * If signature or public key algorithm is unrecognized print out its
13231 OID rather that just UNKNOWN.
13235 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13236 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13237 ID to be generated from the issuer certificate alone which can then be
13238 passed to OCSP_id_issuer_cmp().
13242 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13243 ASN1 modules to export functions returning ASN1_ITEM pointers
13244 instead of the ASN1_ITEM structures themselves. This adds several
13245 new macros which allow the underlying ASN1 function/structure to
13246 be accessed transparently. As a result code should not use ASN1_ITEM
13247 references directly (such as &X509_it) but instead use the relevant
13248 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13249 use of the new ASN1 code on platforms where exporting structures
13250 is problematical (for example in shared libraries) but exporting
13251 functions returning pointers to structures is not.
13255 * Add support for overriding the generation of SSL/TLS session IDs.
13256 These callbacks can be registered either in an SSL_CTX or per SSL.
13257 The purpose of this is to allow applications to control, if they wish,
13258 the arbitrary values chosen for use as session IDs, particularly as it
13259 can be useful for session caching in multiple-server environments. A
13260 command-line switch for testing this (and any client code that wishes
13261 to use such a feature) has been added to "s_server".
13263 *Geoff Thorpe, Lutz Jaenicke*
13265 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13266 of the form `#if defined(...) || defined(...) || ...` and
13267 `#if !defined(...) && !defined(...) && ...`. This also avoids
13268 the growing number of special cases it was previously handling.
13272 * Make all configuration macros available for application by making
13273 sure they are available in opensslconf.h, by giving them names starting
13274 with `OPENSSL_` to avoid conflicts with other packages and by making
13275 sure e_os2.h will cover all platform-specific cases together with
13277 Additionally, it is now possible to define configuration/platform-
13278 specific names (called "system identities"). In the C code, these
13279 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13280 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13281 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13286 * New option -set_serial to 'req' and 'x509' this allows the serial
13287 number to use to be specified on the command line. Previously self
13288 signed certificates were hard coded with serial number 0 and the
13289 CA options of 'x509' had to use a serial number in a file which was
13294 * New options to 'ca' utility to support V2 CRL entry extensions.
13295 Currently CRL reason, invalidity date and hold instruction are
13296 supported. Add new CRL extensions to V3 code and some new objects.
13300 * New function EVP_CIPHER_CTX_set_padding() this is used to
13301 disable standard block padding (aka PKCS#5 padding) in the EVP
13302 API, which was previously mandatory. This means that the data is
13303 not padded in any way and so the total length much be a multiple
13304 of the block size, otherwise an error occurs.
13308 * Initial (incomplete) OCSP SSL support.
13312 * New function OCSP_parse_url(). This splits up a URL into its host,
13313 port and path components: primarily to parse OCSP URLs. New -url
13314 option to ocsp utility.
13318 * New nonce behavior. The return value of OCSP_check_nonce() now
13319 reflects the various checks performed. Applications can decide
13320 whether to tolerate certain situations such as an absent nonce
13321 in a response when one was present in a request: the ocsp application
13322 just prints out a warning. New function OCSP_add1_basic_nonce()
13323 this is to allow responders to include a nonce in a response even if
13324 the request is nonce-less.
13328 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13329 skipped when using openssl x509 multiple times on a single input file,
13330 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13334 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13335 set string type: to handle setting ASN1_TIME structures. Fix ca
13336 utility to correctly initialize revocation date of CRLs.
13340 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13341 the clients preferred ciphersuites and rather use its own preferences.
13342 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13343 Internet Explorer by ensuring unchanged hash method during stepup.
13344 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13348 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13349 to aes and add a new 'exist' option to print out symbols that don't
13354 * Additional options to ocsp utility to allow flags to be set and
13355 additional certificates supplied.
13359 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13360 OCSP client a number of certificate to only verify the response
13365 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13366 handle the new API. Currently only ECB, CBC modes supported. Add new
13369 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13370 Encryption Standard (AES) Ciphersuites for Transport Layer
13371 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13372 not enabled by default and were not part of the "ALL" ciphersuite
13373 alias because they were not yet official; they could be
13374 explicitly requested by specifying the "AESdraft" ciphersuite
13375 group alias. In the final release of OpenSSL 0.9.7, the group
13376 alias is called "AES" and is part of "ALL".)
13378 *Ben Laurie, Steve Henson, Bodo Moeller*
13380 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13381 request to response.
13385 * Functions for OCSP responders. OCSP_request_onereq_count(),
13386 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13387 extract information from a certificate request. OCSP_response_create()
13388 creates a response and optionally adds a basic response structure.
13389 OCSP_basic_add1_status() adds a complete single response to a basic
13390 response and returns the OCSP_SINGLERESP structure just added (to allow
13391 extensions to be included for example). OCSP_basic_add1_cert() adds a
13392 certificate to a basic response and OCSP_basic_sign() signs a basic
13393 response with various flags. New helper functions ASN1_TIME_check()
13394 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13395 (converts ASN1_TIME to GeneralizedTime).
13399 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13400 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13401 structure from a certificate. X509_pubkey_digest() digests the public_key
13402 contents: this is used in various key identifiers.
13406 * Make sk_sort() tolerate a NULL argument.
13408 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13410 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13411 passed by the function are trusted implicitly. If any of them signed the
13412 response then it is assumed to be valid and is not verified.
13416 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13417 to data. This was previously part of the PKCS7 ASN1 code. This
13418 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13419 *Steve Henson, reported by Kenneth R. Robinette
13420 <support@securenetterm.com>*
13422 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13423 routines: without these tracing memory leaks is very painful.
13424 Fix leaks in PKCS12 and PKCS7 routines.
13428 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13429 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13430 effectively meant GeneralizedTime would never be used. Now it
13431 is initialised to -1 but X509_time_adj() now has to check the value
13432 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13433 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13434 *Steve Henson, reported by Kenneth R. Robinette
13435 <support@securenetterm.com>*
13437 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13438 result in a zero length in the ASN1_INTEGER structure which was
13439 not consistent with the structure when d2i_ASN1_INTEGER() was used
13440 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13441 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13442 where it did not print out a minus for negative ASN1_INTEGER.
13446 * Add summary printout to ocsp utility. The various functions which
13447 convert status values to strings have been renamed to:
13448 OCSP_response_status_str(), OCSP_cert_status_str() and
13449 OCSP_crl_reason_str() and are no longer static. New options
13450 to verify nonce values and to disable verification. OCSP response
13451 printout format cleaned up.
13455 * Add additional OCSP certificate checks. These are those specified
13456 in RFC2560. This consists of two separate checks: the CA of the
13457 certificate being checked must either be the OCSP signer certificate
13458 or the issuer of the OCSP signer certificate. In the latter case the
13459 OCSP signer certificate must contain the OCSP signing extended key
13460 usage. This check is performed by attempting to match the OCSP
13461 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13462 in the OCSP_CERTID structures of the response.
13466 * Initial OCSP certificate verification added to OCSP_basic_verify()
13467 and related routines. This uses the standard OpenSSL certificate
13468 verify routines to perform initial checks (just CA validity) and
13469 to obtain the certificate chain. Then additional checks will be
13470 performed on the chain. Currently the root CA is checked to see
13471 if it is explicitly trusted for OCSP signing. This is used to set
13472 a root CA as a global signing root: that is any certificate that
13473 chains to that CA is an acceptable OCSP signing certificate.
13477 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13478 extensions from a separate configuration file.
13479 As when reading extensions from the main configuration file,
13480 the '-extensions ...' option may be used for specifying the
13483 *Massimiliano Pala <madwolf@comune.modena.it>*
13485 * New OCSP utility. Allows OCSP requests to be generated or
13486 read. The request can be sent to a responder and the output
13487 parsed, outputted or printed in text form. Not complete yet:
13488 still needs to check the OCSP response validity.
13492 * New subcommands for 'openssl ca':
13493 `openssl ca -status <serial>` prints the status of the cert with
13494 the given serial number (according to the index file).
13495 `openssl ca -updatedb` updates the expiry status of certificates
13498 *Massimiliano Pala <madwolf@comune.modena.it>*
13500 * New '-newreq-nodes' command option to CA.pl. This is like
13501 '-newreq', but calls 'openssl req' with the '-nodes' option
13502 so that the resulting key is not encrypted.
13504 *Damien Miller <djm@mindrot.org>*
13506 * New configuration for the GNU Hurd.
13508 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13510 * Initial code to implement OCSP basic response verify. This
13511 is currently incomplete. Currently just finds the signer's
13512 certificate and verifies the signature on the response.
13516 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13517 value of OPENSSLDIR. This is available via the new '-d' option
13518 to 'openssl version', and is also included in 'openssl version -a'.
13522 * Allowing defining memory allocation callbacks that will be given
13523 file name and line number information in additional arguments
13524 (a `const char*` and an int). The basic functionality remains, as
13525 well as the original possibility to just replace malloc(),
13526 realloc() and free() by functions that do not know about these
13527 additional arguments. To register and find out the current
13528 settings for extended allocation functions, the following
13529 functions are provided:
13531 CRYPTO_set_mem_ex_functions
13532 CRYPTO_set_locked_mem_ex_functions
13533 CRYPTO_get_mem_ex_functions
13534 CRYPTO_get_locked_mem_ex_functions
13536 These work the same way as CRYPTO_set_mem_functions and friends.
13537 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13538 extended allocation function is enabled.
13539 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13540 a conventional allocation function is enabled.
13542 *Richard Levitte, Bodo Moeller*
13544 * Finish off removing the remaining LHASH function pointer casts.
13545 There should no longer be any prototype-casting required when using
13546 the LHASH abstraction, and any casts that remain are "bugs". See
13547 the callback types and macros at the head of lhash.h for details
13548 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13552 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13553 If /dev/[u]random devices are not available or do not return enough
13554 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13556 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13557 /etc/entropy will be queried once each in this sequence, querying stops
13558 when enough entropy was collected without querying more sockets.
13562 * Change the Unix RAND_poll() variant to be able to poll several
13563 random devices, as specified by DEVRANDOM, until a sufficient amount
13564 of data has been collected. We spend at most 10 ms on each file
13565 (select timeout) and read in non-blocking mode. DEVRANDOM now
13566 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13567 (previously it was just the string "/dev/urandom"), so on typical
13568 platforms the 10 ms delay will never occur.
13569 Also separate out the Unix variant to its own file, rand_unix.c.
13570 For VMS, there's a currently-empty rand_vms.c.
13574 * Move OCSP client related routines to ocsp_cl.c. These
13575 provide utility functions which an application needing
13576 to issue a request to an OCSP responder and analyse the
13577 response will typically need: as opposed to those which an
13578 OCSP responder itself would need which will be added later.
13580 OCSP_request_sign() signs an OCSP request with an API similar
13581 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13582 response. OCSP_response_get1_basic() extracts basic response
13583 from response. OCSP_resp_find_status(): finds and extracts status
13584 information from an OCSP_CERTID structure (which will be created
13585 when the request structure is built). These are built from lower
13586 level functions which work on OCSP_SINGLERESP structures but
13587 won't normally be used unless the application wishes to examine
13588 extensions in the OCSP response for example.
13590 Replace nonce routines with a pair of functions.
13591 OCSP_request_add1_nonce() adds a nonce value and optionally
13592 generates a random value. OCSP_check_nonce() checks the
13593 validity of the nonce in an OCSP response.
13597 * Change function OCSP_request_add() to OCSP_request_add0_id().
13598 This doesn't copy the supplied OCSP_CERTID and avoids the
13599 need to free up the newly created id. Change return type
13600 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13601 This can then be used to add extensions to the request.
13602 Deleted OCSP_request_new(), since most of its functionality
13603 is now in OCSP_REQUEST_new() (and the case insensitive name
13604 clash) apart from the ability to set the request name which
13605 will be added elsewhere.
13609 * Update OCSP API. Remove obsolete extensions argument from
13610 various functions. Extensions are now handled using the new
13611 OCSP extension code. New simple OCSP HTTP function which
13612 can be used to send requests and parse the response.
13616 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13617 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13618 uses the special reorder version of SET OF to sort the attributes
13619 and reorder them to match the encoded order. This resolves a long
13620 standing problem: a verify on a PKCS7 structure just after signing
13621 it used to fail because the attribute order did not match the
13622 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13623 it uses the received order. This is necessary to tolerate some broken
13624 software that does not order SET OF. This is handled by encoding
13625 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13626 to produce the required SET OF.
13630 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13631 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13632 files to get correct declarations of the ASN.1 item variables.
13636 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13637 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13638 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13639 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13640 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13641 ASN1_ITEM and no wrapper functions.
13645 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13646 replace the old function pointer based I/O routines. Change most of
13647 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13651 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13652 lines, recognize more "algorithms" that can be deselected, and make
13653 it complain about algorithm deselection that isn't recognised.
13657 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13658 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13659 to use new functions. Add NO_ASN1_OLD which can be set to remove
13660 some old style ASN1 functions: this can be used to determine if old
13661 code will still work when these eventually go away.
13665 * New extension functions for OCSP structures, these follow the
13666 same conventions as certificates and CRLs.
13670 * New function X509V3_add1_i2d(). This automatically encodes and
13671 adds an extension. Its behaviour can be customised with various
13672 flags to append, replace or delete. Various wrappers added for
13673 certificates and CRLs.
13677 * Fix to avoid calling the underlying ASN1 print routine when
13678 an extension cannot be parsed. Correct a typo in the
13679 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13683 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13684 entries for variables.
13688 * Add functionality to `apps/openssl.c` for detecting locking
13689 problems: As the program is single-threaded, all we have
13690 to do is register a locking callback using an array for
13691 storing which locks are currently held by the program.
13695 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13696 SSL_get_ex_data_X509_STORE_idx(), which is used in
13697 ssl_verify_cert_chain() and thus can be called at any time
13698 during TLS/SSL handshakes so that thread-safety is essential.
13699 Unfortunately, the ex_data design is not at all suited
13700 for multi-threaded use, so it probably should be abolished.
13704 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13706 *Broadcom, tweaked and integrated by Geoff Thorpe*
13708 * Move common extension printing code to new function
13709 X509V3_print_extensions(). Reorganise OCSP print routines and
13710 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13714 * New function X509_signature_print() to remove duplication in some
13719 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13720 set (this was treated exactly the same as SET OF previously). This
13721 is used to reorder the STACK representing the structure to match the
13722 encoding. This will be used to get round a problem where a PKCS7
13723 structure which was signed could not be verified because the STACK
13724 order did not reflect the encoded order.
13728 * Reimplement the OCSP ASN1 module using the new code.
13732 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13733 for its ASN1 operations. The old style function pointers still exist
13734 for now but they will eventually go away.
13738 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13739 completely replaces the old ASN1 functionality with a table driven
13740 encoder and decoder which interprets an ASN1_ITEM structure describing
13741 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13742 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13743 has also been converted to the new form.
13747 * Change BN_mod_exp_recp so that negative moduli are tolerated
13748 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13749 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13750 for negative moduli.
13754 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13755 of not touching the result's sign bit.
13759 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13764 * Changed the LHASH code to use prototypes for callbacks, and created
13765 macros to declare and implement thin (optionally static) functions
13766 that provide type-safety and avoid function pointer casting for the
13767 type-specific callbacks.
13771 * Added Kerberos Cipher Suites to be used with TLS, as written in
13773 *Veers Staats <staatsvr@asc.hpc.mil>,
13774 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13776 * Reformat the FAQ so the different questions and answers can be divided
13777 in sections depending on the subject.
13781 * Have the zlib compression code load ZLIB.DLL dynamically under
13786 * New function BN_mod_sqrt for computing square roots modulo a prime
13787 (using the probabilistic Tonelli-Shanks algorithm unless
13788 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13789 be handled deterministically).
13791 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13793 * Make BN_mod_inverse faster by explicitly handling small quotients
13794 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13795 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13799 * New function BN_kronecker.
13803 * Fix BN_gcd so that it works on negative inputs; the result is
13804 positive unless both parameters are zero.
13805 Previously something reasonably close to an infinite loop was
13806 possible because numbers could be growing instead of shrinking
13807 in the implementation of Euclid's algorithm.
13811 * Fix BN_is_word() and BN_is_one() macros to take into account the
13812 sign of the number in question.
13814 Fix BN_is_word(a,w) to work correctly for w == 0.
13816 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13817 because its test if the absolute value of 'a' equals 'w'.
13818 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13819 it exists mostly for use in the implementations of BN_is_zero(),
13820 BN_is_one(), and BN_is_word().
13824 * New function BN_swap.
13828 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13829 the exponentiation functions are more likely to produce reasonable
13830 results on negative inputs.
13834 * Change BN_mod_mul so that the result is always non-negative.
13835 Previously, it could be negative if one of the factors was negative;
13836 I don't think anyone really wanted that behaviour.
13840 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13841 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13842 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13843 and add new functions:
13852 BN_mod_lshift1_quick
13854 BN_mod_lshift_quick
13856 These functions always generate non-negative results.
13858 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13859 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13861 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13862 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13863 be reduced modulo `m`.
13865 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13868 The following entry accidentally appeared in the CHANGES file
13869 distributed with OpenSSL 0.9.7. The modifications described in
13870 it do *not* apply to OpenSSL 0.9.7.
13872 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13873 was actually never needed) and in BN_mul(). The removal in BN_mul()
13874 required a small change in bn_mul_part_recursive() and the addition
13875 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13876 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13877 bn_sub_words() and bn_add_words() except they take arrays with
13883 * In 'openssl passwd', verify passwords read from the terminal
13884 unless the '-salt' option is used (which usually means that
13885 verification would just waste user's time since the resulting
13886 hash is going to be compared with some given password hash)
13887 or the new '-noverify' option is used.
13889 This is an incompatible change, but it does not affect
13890 non-interactive use of 'openssl passwd' (passwords on the command
13891 line, '-stdin' option, '-in ...' option) and thus should not
13892 cause any problems.
13896 * Remove all references to RSAref, since there's no more need for it.
13900 * Make DSO load along a path given through an environment variable
13901 (SHLIB_PATH) with shl_load().
13905 * Constify the ENGINE code as a result of BIGNUM constification.
13906 Also constify the RSA code and most things related to it. In a
13907 few places, most notable in the depth of the ASN.1 code, ugly
13908 casts back to non-const were required (to be solved at a later
13913 * Make it so the openssl application has all engines loaded by default.
13917 * Constify the BIGNUM routines a little more.
13921 * Add the following functions:
13923 ENGINE_load_cswift()
13925 ENGINE_load_atalla()
13926 ENGINE_load_nuron()
13927 ENGINE_load_builtin_engines()
13929 That way, an application can itself choose if external engines that
13930 are built-in in OpenSSL shall ever be used or not. The benefit is
13931 that applications won't have to be linked with libdl or other dso
13932 libraries unless it's really needed.
13934 Changed 'openssl engine' to load all engines on demand.
13935 Changed the engine header files to avoid the duplication of some
13936 declarations (they differed!).
13940 * 'openssl engine' can now list capabilities.
13944 * Better error reporting in 'openssl engine'.
13948 * Never call load_dh_param(NULL) in s_server.
13952 * Add engine application. It can currently list engines by name and
13953 identity, and test if they are actually available.
13957 * Improve RPM specification file by forcing symbolic linking and making
13958 sure the installed documentation is also owned by root.root.
13960 *Damien Miller <djm@mindrot.org>*
13962 * Give the OpenSSL applications more possibilities to make use of
13963 keys (public as well as private) handled by engines.
13967 * Add OCSP code that comes from CertCo.
13971 * Add VMS support for the Rijndael code.
13975 * Added untested support for Nuron crypto accelerator.
13979 * Add support for external cryptographic devices. This code was
13980 previously distributed separately as the "engine" branch.
13982 *Geoff Thorpe, Richard Levitte*
13984 * Rework the filename-translation in the DSO code. It is now possible to
13985 have far greater control over how a "name" is turned into a filename
13986 depending on the operating environment and any oddities about the
13987 different shared library filenames on each system.
13991 * Support threads on FreeBSD-elf in Configure.
13995 * Fix for SHA1 assembly problem with MASM: it produces
13996 warnings about corrupt line number information when assembling
13997 with debugging information. This is caused by the overlapping
14000 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14003 NCONF_get_number() has no error checking at all. As a replacement,
14004 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14005 promoted strongly. The old NCONF_get_number is kept around for
14006 binary backward compatibility.
14007 Make it possible for methods to load from something other than a BIO,
14008 by providing a function pointer that is given a name instead of a BIO.
14009 For example, this could be used to load configuration data from an
14014 * Fix for non blocking accept BIOs. Added new I/O special reason
14015 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14016 with non blocking I/O was not possible because no retry code was
14017 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14022 * Added the beginnings of Rijndael support.
14026 * Fix for bug in DirectoryString mask setting. Add support for
14027 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14028 to allow certificate printing to more controllable, additional
14029 'certopt' option to 'x509' to allow new printing options to be
14034 * Clean old EAY MD5 hack from e_os.h.
14038 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14040 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14041 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14043 *Joe Orton, Steve Henson*
14045 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14047 * Fix additional bug revealed by the NISCC test suite:
14049 Stop bug triggering large recursion when presented with
14050 certain ASN.1 tags ([CVE-2003-0851])
14054 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14056 * Fix various bugs revealed by running the NISCC test suite:
14058 Stop out of bounds reads in the ASN1 code when presented with
14059 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14061 If verify callback ignores invalid public key errors don't try to check
14062 certificate signature with the NULL public key.
14066 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14067 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14072 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14073 extra data after the compression methods not only for TLS 1.0
14074 but also for SSL 3.0 (as required by the specification).
14076 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14078 * Change X509_certificate_type() to mark the key as exported/exportable
14079 when it's 512 *bits* long, not 512 bytes.
14083 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14085 * Countermeasure against the Klima-Pokorny-Rosa extension of
14086 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14087 a protocol version number mismatch like a decryption error
14088 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14092 * Turn on RSA blinding by default in the default implementation
14093 to avoid a timing attack. Applications that don't want it can call
14094 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14095 They would be ill-advised to do so in most cases.
14097 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14099 * Change RSA blinding code so that it works when the PRNG is not
14100 seeded (in this case, the secret RSA exponent is abused as
14101 an unpredictable seed -- if it is not unpredictable, there
14102 is no point in blinding anyway). Make RSA blinding thread-safe
14103 by remembering the creator's thread ID in rsa->blinding and
14104 having all other threads use local one-time blinding factors
14105 (this requires more computation than sharing rsa->blinding, but
14106 avoids excessive locking; and if an RSA object is not shared
14107 between threads, blinding will still be very fast).
14111 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14113 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14114 via timing by performing a MAC computation even if incorrect
14115 block cipher padding has been found. This is a countermeasure
14116 against active attacks where the attacker has to distinguish
14117 between bad padding and a MAC verification error. ([CVE-2003-0078])
14119 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14120 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14121 Martin Vuagnoux (EPFL, Ilion)*
14123 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14125 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14126 memory from its contents. This is done with a counter that will
14127 place alternating values in each byte. This can be used to solve
14128 two issues: 1) the removal of calls to memset() by highly optimizing
14129 compilers, and 2) cleansing with other values than 0, since those can
14130 be read through on certain media, for example a swap space on disk.
14134 * Bugfix: client side session caching did not work with external caching,
14135 because the session->cipher setting was not restored when reloading
14136 from the external cache. This problem was masked, when
14137 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14138 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14142 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14143 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14145 *Zeev Lieber <zeev-l@yahoo.com>*
14147 * Undo an undocumented change introduced in 0.9.6e which caused
14148 repeated calls to OpenSSL_add_all_ciphers() and
14149 OpenSSL_add_all_digests() to be ignored, even after calling
14154 * Change the default configuration reader to deal with last line not
14155 being properly terminated.
14159 * Change X509_NAME_cmp() so it applies the special rules on handling
14160 DN values that are of type PrintableString, as well as RDNs of type
14161 emailAddress where the value has the type ia5String.
14163 *stefank@valicert.com via Richard Levitte*
14165 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14166 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14167 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14168 the bitwise-OR of the two for use by the majority of applications
14169 wanting this behaviour, and update the docs. The documented
14170 behaviour and actual behaviour were inconsistent and had been
14171 changing anyway, so this is more a bug-fix than a behavioural
14174 *Geoff Thorpe, diagnosed by Nadav Har'El*
14176 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14177 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14181 * Fix initialization code race conditions in
14182 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14183 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14184 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14185 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14186 ssl2_get_cipher_by_char(),
14187 ssl3_get_cipher_by_char().
14189 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14191 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14192 the cached sessions are flushed, as the remove_cb() might use ex_data
14193 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14194 (see [openssl.org #212]).
14196 *Geoff Thorpe, Lutz Jaenicke*
14198 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14199 length, instead of the encoding length to d2i_ASN1_OBJECT.
14203 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14205 * [In 0.9.6g-engine release:]
14206 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14208 *Lynn Gazis <lgazis@rainbow.com>*
14210 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14212 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14213 and get fix the header length calculation.
14214 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14215 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14217 * Use proper error handling instead of 'assertions' in buffer
14218 overflow checks added in 0.9.6e. This prevents DoS (the
14219 assertions could call abort()).
14221 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14223 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14225 * Add various sanity checks to asn1_get_length() to reject
14226 the ASN1 length bytes if they exceed sizeof(long), will appear
14227 negative or the content length exceeds the length of the
14230 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14232 * Fix cipher selection routines: ciphers without encryption had no flags
14233 for the cipher strength set and where therefore not handled correctly
14234 by the selection routines (PR #130).
14238 * Fix EVP_dsa_sha macro.
14243 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14244 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14245 that was added in OpenSSL 0.9.6d.
14247 As the countermeasure turned out to be incompatible with some
14248 broken SSL implementations, the new option is part of SSL_OP_ALL.
14249 SSL_OP_ALL is usually employed when compatibility with weird SSL
14250 implementations is desired (e.g. '-bugs' option to 's_client' and
14251 's_server'), so the new option is automatically set in many
14256 * Changes in security patch:
14258 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14259 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14260 Air Force Materiel Command, USAF, under agreement number
14263 * Add various sanity checks to asn1_get_length() to reject
14264 the ASN1 length bytes if they exceed sizeof(long), will appear
14265 negative or the content length exceeds the length of the
14266 supplied buffer. ([CVE-2002-0659])
14268 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14270 * Assertions for various potential buffer overflows, not known to
14271 happen in practice.
14273 *Ben Laurie (CHATS)*
14275 * Various temporary buffers to hold ASCII versions of integers were
14276 too small for 64 bit platforms. ([CVE-2002-0655])
14277 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14279 * Remote buffer overflow in SSL3 protocol - an attacker could
14280 supply an oversized session ID to a client. ([CVE-2002-0656])
14282 *Ben Laurie (CHATS)*
14284 * Remote buffer overflow in SSL2 protocol - an attacker could
14285 supply an oversized client master key. ([CVE-2002-0656])
14287 *Ben Laurie (CHATS)*
14289 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14291 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14292 encoded as NULL) with id-dsa-with-sha1.
14294 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14296 * Check various `X509_...()` return values in `apps/req.c`.
14298 *Nils Larsch <nla@trustcenter.de>*
14300 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14301 an end-of-file condition would erroneously be flagged, when the CRLF
14302 was just at the end of a processed block. The bug was discovered when
14303 processing data through a buffering memory BIO handing the data to a
14304 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14305 <ptsekov@syntrex.com> and Nedelcho Stanev.
14309 * Implement a countermeasure against a vulnerability recently found
14310 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14311 before application data chunks to avoid the use of known IVs
14312 with data potentially chosen by the attacker.
14316 * Fix length checks in ssl3_get_client_hello().
14320 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14321 to prevent ssl3_read_internal() from incorrectly assuming that
14322 ssl3_read_bytes() found application data while handshake
14323 processing was enabled when in fact s->s3->in_read_app_data was
14324 merely automatically cleared during the initial handshake.
14326 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14328 * Fix object definitions for Private and Enterprise: they were not
14329 recognized in their shortname (=lowercase) representation. Extend
14330 obj_dat.pl to issue an error when using undefined keywords instead
14331 of silently ignoring the problem (Svenning Sorensen
14332 <sss@sss.dnsalias.net>).
14336 * Fix DH_generate_parameters() so that it works for 'non-standard'
14337 generators, i.e. generators other than 2 and 5. (Previously, the
14338 code did not properly initialise the 'add' and 'rem' values to
14339 BN_generate_prime().)
14341 In the new general case, we do not insist that 'generator' is
14342 actually a primitive root: This requirement is rather pointless;
14343 a generator of the order-q subgroup is just as good, if not
14348 * Map new X509 verification errors to alerts. Discovered and submitted by
14349 Tom Wu <tom@arcot.com>.
14353 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14354 returning non-zero before the data has been completely received
14355 when using non-blocking I/O.
14357 *Bodo Moeller; problem pointed out by John Hughes*
14359 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14361 *Ben Laurie, Lutz Jaenicke*
14363 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14364 Yoram Zahavi <YoramZ@gilian.com>).
14368 * Add information about CygWin 1.3 and on, and preserve proper
14369 configuration for the versions before that.
14371 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14373 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14374 check whether we deal with a copy of a session and do not delete from
14375 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14376 <izhar@checkpoint.com>.
14380 * Do not store session data into the internal session cache, if it
14381 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14382 flag is set). Proposed by Aslam <aslam@funk.com>.
14386 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14391 * [In 0.9.6d-engine release:]
14392 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14394 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14396 * Add the configuration target linux-s390x.
14398 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14400 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14401 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14402 variable as an indication that a ClientHello message has been
14403 received. As the flag value will be lost between multiple
14404 invocations of ssl3_accept when using non-blocking I/O, the
14405 function may not be aware that a handshake has actually taken
14406 place, thus preventing a new session from being added to the
14409 To avoid this problem, we now set s->new_session to 2 instead of
14410 using a local variable.
14412 *Lutz Jaenicke, Bodo Moeller*
14414 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14415 if the SSL_R_LENGTH_MISMATCH error is detected.
14417 *Geoff Thorpe, Bodo Moeller*
14419 * New 'shared_ldflag' column in Configure platform table.
14423 * Fix EVP_CIPHER_mode macro.
14425 *"Dan S. Camper" <dan@bti.net>*
14427 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14428 type, we must throw them away by setting rr->length to 0.
14430 *D P Chang <dpc@qualys.com>*
14432 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14434 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14435 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14436 worked incorrectly for those cases where range = `10..._2` and
14437 `3*range` is two bits longer than range.)
14441 * Only add signing time to PKCS7 structures if it is not already
14446 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14447 OBJ_ld_ce should be OBJ_id_ce.
14448 Also some ip-pda OIDs in crypto/objects/objects.txt were
14449 incorrect (cf. RFC 3039).
14451 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14453 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14454 returns early because it has nothing to do.
14456 *Andy Schneider <andy.schneider@bjss.co.uk>*
14458 * [In 0.9.6c-engine release:]
14459 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14461 *Andy Schneider <andy.schneider@bjss.co.uk>*
14463 * [In 0.9.6c-engine release:]
14464 Add support for Cryptographic Appliance's keyserver technology.
14465 (Use engine 'keyclient')
14467 *Cryptographic Appliances and Geoff Thorpe*
14469 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14470 is called via tools/c89.sh because arguments have to be
14471 rearranged (all '-L' options must appear before the first object
14474 *Richard Shapiro <rshapiro@abinitio.com>*
14476 * [In 0.9.6c-engine release:]
14477 Add support for Broadcom crypto accelerator cards, backported
14480 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14482 * [In 0.9.6c-engine release:]
14483 Add support for SureWare crypto accelerator cards from
14484 Baltimore Technologies. (Use engine 'sureware')
14486 *Baltimore Technologies and Mark Cox*
14488 * [In 0.9.6c-engine release:]
14489 Add support for crypto accelerator cards from Accelerated
14490 Encryption Processing, www.aep.ie. (Use engine 'aep')
14492 *AEP Inc. and Mark Cox*
14494 * Add a configuration entry for gcc on UnixWare.
14496 *Gary Benson <gbenson@redhat.com>*
14498 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14499 messages are stored in a single piece (fixed-length part and
14500 variable-length part combined) and fix various bugs found on the way.
14504 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14505 instead. BIO_gethostbyname() does not know what timeouts are
14506 appropriate, so entries would stay in cache even when they have
14508 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14510 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14511 faced with a pathologically small ClientHello fragment that does
14512 not contain client_version: Instead of aborting with an error,
14513 simply choose the highest available protocol version (i.e.,
14514 TLS 1.0 unless it is disabled). In practice, ClientHello
14515 messages are never sent like this, but this change gives us
14516 strictly correct behaviour at least for TLS.
14520 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14521 never resets s->method to s->ctx->method when called from within
14522 one of the SSL handshake functions.
14524 *Bodo Moeller; problem pointed out by Niko Baric*
14526 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14527 (sent using the client's version number) if client_version is
14528 smaller than the protocol version in use. Also change
14529 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14530 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14531 the client will at least see that alert.
14535 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14540 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14541 client receives HelloRequest while in a handshake.
14543 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14545 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14546 should end in 'break', not 'goto end' which circumvents various
14547 cleanups done in state SSL_ST_OK. But session related stuff
14548 must be disabled for SSL_ST_OK in the case that we just sent a
14551 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14552 before just sending a HelloRequest.
14554 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14556 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14557 reveal whether illegal block cipher padding was found or a MAC
14558 verification error occurred. (Neither SSLerr() codes nor alerts
14559 are directly visible to potential attackers, but the information
14560 may leak via logfiles.)
14562 Similar changes are not required for the SSL 2.0 implementation
14563 because the number of padding bytes is sent in clear for SSL 2.0,
14564 and the extra bytes are just ignored. However ssl/s2_pkt.c
14565 failed to verify that the purported number of padding bytes is in
14570 * Add OpenUNIX-8 support including shared libraries
14571 (Boyd Lynn Gerber <gerberb@zenez.com>).
14575 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14576 'wristwatch attack' using huge encoding parameters (cf.
14577 James H. Manger's CRYPTO 2001 paper). Note that the
14578 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14579 encoding parameters and hence was not vulnerable.
14583 * BN_sqr() bug fix.
14585 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14587 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14588 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14589 followed by modular reduction.
14591 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14593 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14594 equivalent based on BN_pseudo_rand() instead of BN_rand().
14598 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14599 This function was broken, as the check for a new client hello message
14600 to handle SGC did not allow these large messages.
14601 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14605 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14609 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14610 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14614 * Rework the configuration and shared library support for Tru64 Unix.
14615 The configuration part makes use of modern compiler features and
14616 still retains old compiler behavior for those that run older versions
14617 of the OS. The shared library support part includes a variant that
14618 uses the RPATH feature, and is available through the special
14619 configuration target "alpha-cc-rpath", which will never be selected
14622 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14624 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14625 with the same message size as in ssl3_get_certificate_request().
14626 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14627 messages might inadvertently be reject as too long.
14629 *Petr Lampa <lampa@fee.vutbr.cz>*
14631 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14635 * Modified SSL library such that the verify_callback that has been set
14636 specifically for an SSL object with SSL_set_verify() is actually being
14637 used. Before the change, a verify_callback set with this function was
14638 ignored and the verify_callback() set in the SSL_CTX at the time of
14639 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14640 to allow the necessary settings.
14644 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14645 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14646 done automatically (in contradiction to the requirements of the C
14647 standard). This made problems when used from OpenSSH.
14651 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14652 dh->length and always used
14654 BN_rand_range(priv_key, dh->p).
14656 BN_rand_range() is not necessary for Diffie-Hellman, and this
14657 specific range makes Diffie-Hellman unnecessarily inefficient if
14658 dh->length (recommended exponent length) is much smaller than the
14659 length of dh->p. We could use BN_rand_range() if the order of
14660 the subgroup was stored in the DH structure, but we only have
14665 BN_rand(priv_key, l, ...)
14667 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14674 RSA_eay_public_encrypt
14675 RSA_eay_private_decrypt
14676 RSA_eay_private_encrypt (signing)
14677 RSA_eay_public_decrypt (signature verification)
14679 (default implementations for RSA_public_encrypt,
14680 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14681 always reject numbers >= n.
14685 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14686 to synchronize access to 'locking_thread'. This is necessary on
14687 systems where access to 'locking_thread' (an 'unsigned long'
14688 variable) is not atomic.
14692 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14693 *before* setting the 'crypto_lock_rand' flag. The previous code had
14694 a race condition if 0 is a valid thread ID.
14696 *Travis Vitek <vitek@roguewave.com>*
14698 * Add support for shared libraries under Irix.
14700 *Albert Chin-A-Young <china@thewrittenword.com>*
14702 * Add configuration option to build on Linux on both big-endian and
14703 little-endian MIPS.
14705 *Ralf Baechle <ralf@uni-koblenz.de>*
14707 * Add the possibility to create shared libraries on HP-UX.
14711 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14713 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14714 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14715 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14716 PRNG state recovery was possible based on the output of
14717 one PRNG request appropriately sized to gain knowledge on
14718 'md' followed by enough consecutive 1-byte PRNG requests
14719 to traverse all of 'state'.
14721 1. When updating 'md_local' (the current thread's copy of 'md')
14722 during PRNG output generation, hash all of the previous
14723 'md_local' value, not just the half used for PRNG output.
14725 2. Make the number of bytes from 'state' included into the hash
14726 independent from the number of PRNG bytes requested.
14728 The first measure alone would be sufficient to avoid
14729 Markku-Juhani's attack. (Actually it had never occurred
14730 to me that the half of 'md_local' used for chaining was the
14731 half from which PRNG output bytes were taken -- I had always
14732 assumed that the secret half would be used.) The second
14733 measure makes sure that additional data from 'state' is never
14734 mixed into 'md_local' in small portions; this heuristically
14735 further strengthens the PRNG.
14739 * Fix crypto/bn/asm/mips3.s.
14743 * When only the key is given to "enc", the IV is undefined. Print out
14744 an error message in this case.
14748 * Handle special case when X509_NAME is empty in X509 printing routines.
14752 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14753 positive and less than q.
14757 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14758 used: it isn't thread safe and the add_lock_callback should handle
14761 *Paul Rose <Paul.Rose@bridge.com>*
14763 * Verify that incoming data obeys the block size in
14764 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14770 *Ulf Möller, Bodo Möller*
14772 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14773 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14774 when fixing the server behaviour for backwards-compatible 'client
14775 hello' messages. (Note that the attack is impractical against
14776 SSL 3.0 and TLS 1.0 anyway because length and version checking
14777 means that the probability of guessing a valid ciphertext is
14778 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14781 Before 0.9.5, the countermeasure (hide the error by generating a
14782 random 'decryption result') did not work properly because
14783 ERR_clear_error() was missing, meaning that SSL_get_error() would
14784 detect the supposedly ignored error.
14786 Both problems are now fixed.
14790 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14791 (previously it was 1024).
14795 * Fix for compatibility mode trust settings: ignore trust settings
14796 unless some valid trust or reject settings are present.
14800 * Fix for blowfish EVP: its a variable length cipher.
14804 * Fix various bugs related to DSA S/MIME verification. Handle missing
14805 parameters in DSA public key structures and return an error in the
14806 DSA routines if parameters are absent.
14810 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14811 in the current directory if neither $RANDFILE nor $HOME was set.
14812 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14813 caused some confusion to Windows users who haven't defined $HOME.
14814 Thus RAND_file_name() is changed again: e_os.h can define a
14815 DEFAULT_HOME, which will be used if $HOME is not set.
14816 For Windows, we use "C:"; on other platforms, we still require
14817 environment variables.
14819 * Move 'if (!initialized) RAND_poll()' into regions protected by
14820 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14821 having multiple threads call RAND_poll() concurrently.
14825 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14826 combination of a flag and a thread ID variable.
14827 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14828 flag), *other* threads can enter ssleay_add_bytes without obeying
14829 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14830 that they do not hold after the first thread unsets add_do_not_lock).
14834 * Change bctest again: '-x' expressions are not available in all
14835 versions of 'test'.
14839 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14841 * Fix a couple of memory leaks in PKCS7_dataDecode()
14843 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14845 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14846 the default extension for executables, if any. Also, make the perl
14847 scripts that use symlink() to test if it really exists and use "cp"
14848 if it doesn't. All this made OpenSSL compilable and installable in
14853 * Fix for asn1_GetSequence() for indefinite length constructed data.
14854 If SEQUENCE is length is indefinite just set c->slen to the total
14855 amount of data available.
14857 *Steve Henson, reported by shige@FreeBSD.org*
14859 *This change does not apply to 0.9.7.*
14861 * Change bctest to avoid here-documents inside command substitution
14862 (workaround for FreeBSD /bin/sh bug).
14863 For compatibility with Ultrix, avoid shell functions (introduced
14864 in the bctest version that searches along $PATH).
14868 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14869 with des_encrypt() defined on some operating systems, like Solaris
14874 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14875 On the Importance of Eliminating Errors in Cryptographic
14876 Computations, J. Cryptology 14 (2001) 2, 101-119,
14877 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14881 * MIPS assembler BIGNUM division bug fix.
14885 * Disabled incorrect Alpha assembler code.
14889 * Fix PKCS#7 decode routines so they correctly update the length
14890 after reading an EOC for the EXPLICIT tag.
14894 *This change does not apply to 0.9.7.*
14896 * Fix bug in PKCS#12 key generation routines. This was triggered
14897 if a 3DES key was generated with a 0 initial byte. Include
14898 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14899 (but broken) behaviour.
14903 * Enhance bctest to search for a working bc along $PATH and print
14906 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14908 * Fix memory leaks in err.c: free err_data string if necessary;
14909 don't write to the wrong index in ERR_set_error_data.
14913 * Implement ssl23_peek (analogous to ssl23_read), which previously
14918 * Replace rdtsc with `_emit` statements for VC++ version 5.
14920 *Jeremy Cooper <jeremy@baymoo.org>*
14922 * Make it possible to reuse SSLv2 sessions.
14926 * In copy_email() check for >= 0 as a return value for
14927 X509_NAME_get_index_by_NID() since 0 is a valid index.
14929 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14931 * Avoid coredump with unsupported or invalid public keys by checking if
14932 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14933 PKCS7_verify() fails with non detached data.
14937 * Don't use getenv in library functions when run as setuid/setgid.
14938 New function OPENSSL_issetugid().
14942 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14943 due to incorrect handling of multi-threading:
14945 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14947 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14949 3. Count how many times MemCheck_off() has been called so that
14950 nested use can be treated correctly. This also avoids
14951 inband-signalling in the previous code (which relied on the
14952 assumption that thread ID 0 is impossible).
14956 * Add "-rand" option also to s_client and s_server.
14960 * Fix CPU detection on Irix 6.x.
14961 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14962 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14964 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14969 *This change does not apply to 0.9.7.*
14971 * Use the cached encoding of an X509_NAME structure rather than
14972 copying it. This is apparently the reason for the libsafe "errors"
14973 but the code is actually correct.
14977 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14978 Bleichenbacher's DSA attack.
14979 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14980 to be set and top=0 forces the highest bit to be set; top=-1 is new
14981 and leaves the highest bit random.
14983 *Ulf Moeller, Bodo Moeller*
14985 * In the `NCONF_...`-based implementations for `CONF_...` queries
14986 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14987 a temporary CONF structure with the data component set to NULL
14988 (which gives segmentation faults in lh_retrieve).
14989 Instead, use NULL for the CONF pointer in CONF_get_string and
14990 CONF_get_number (which may use environment variables) and directly
14991 return NULL from CONF_get_section.
14995 * Fix potential buffer overrun for EBCDIC.
14999 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15000 keyUsage if basicConstraints absent for a CA.
15004 * Make SMIME_write_PKCS7() write mail header values with a format that
15005 is more generally accepted (no spaces before the semicolon), since
15006 some programs can't parse those values properly otherwise. Also make
15007 sure BIO's that break lines after each write do not create invalid
15012 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15013 macros previously used would not encode an empty SEQUENCE OF
15014 and break the signature.
15018 *This change does not apply to 0.9.7.*
15020 * Zero the premaster secret after deriving the master secret in
15025 * Add some EVP_add_digest_alias registrations (as found in
15026 OpenSSL_add_all_digests()) to SSL_library_init()
15027 aka OpenSSL_add_ssl_algorithms(). This provides improved
15028 compatibility with peers using X.509 certificates
15029 with unconventional AlgorithmIdentifier OIDs.
15033 * Fix for Irix with NO_ASM.
15035 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15037 * ./config script fixes.
15039 *Ulf Moeller, Richard Levitte*
15041 * Fix 'openssl passwd -1'.
15045 * Change PKCS12_key_gen_asc() so it can cope with non null
15046 terminated strings whose length is passed in the passlen
15047 parameter, for example from PEM callbacks. This was done
15048 by adding an extra length parameter to asc2uni().
15050 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15052 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15053 call failed, free the DSA structure.
15057 * Fix to uni2asc() to cope with zero length Unicode strings.
15058 These are present in some PKCS#12 files.
15062 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15063 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15064 when writing a 32767 byte record.
15066 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15068 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15069 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15071 (RSA objects have a reference count access to which is protected
15072 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15073 so they are meant to be shared between threads.)
15074 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15075 "Reddie, Steven" <Steven.Reddie@ca.com>*
15077 * Fix a deadlock in CRYPTO_mem_leaks().
15081 * Use better test patterns in bntest.
15085 * rand_win.c fix for Borland C.
15089 * BN_rshift bugfix for n == 0.
15093 * Add a 'bctest' script that checks for some known 'bc' bugs
15094 so that 'make test' does not abort just because 'bc' is broken.
15098 * Store verify_result within SSL_SESSION also for client side to
15099 avoid potential security hole. (Re-used sessions on the client side
15100 always resulted in verify_result==X509_V_OK, not using the original
15101 result of the server certificate verification.)
15105 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15106 SSL3_RT_APPLICATION_DATA, return 0.
15107 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15112 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15113 releases, have been re-implemented by renaming the previous
15114 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15115 and ssl3_read_internal, respectively, and adding 'peek' parameters
15116 to them. The new ssl[23]_{read,peek} functions are calls to
15117 ssl[23]_read_internal with the 'peek' flag set appropriately.
15118 A 'peek' parameter has also been added to ssl3_read_bytes, which
15119 does the actual work for ssl3_read_internal.
15123 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15124 the method-specific "init()" handler. Also clean up ex_data after
15125 calling the method-specific "finish()" handler. Previously, this was
15126 happening the other way round.
15130 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15131 The previous value, 12, was not always sufficient for BN_mod_exp().
15135 * Make sure that shared libraries get the internal name engine with
15136 the full version number and not just 0. This should mark the
15137 shared libraries as not backward compatible. Of course, this should
15138 be changed again when we can guarantee backward binary compatibility.
15142 * Fix typo in get_cert_by_subject() in by_dir.c
15144 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15146 * Rework the system to generate shared libraries:
15148 - Make note of the expected extension for the shared libraries and
15149 if there is a need for symbolic links from for example libcrypto.so.0
15150 to libcrypto.so.0.9.7. There is extended info in Configure for
15153 - Make as few rebuilds of the shared libraries as possible.
15155 - Still avoid linking the OpenSSL programs with the shared libraries.
15157 - When installing, install the shared libraries separately from the
15162 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15164 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15165 and not in SSL_clear because the latter is also used by the
15166 accept/connect functions; previously, the settings made by
15167 SSL_set_read_ahead would be lost during the handshake.
15169 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15171 * Correct util/mkdef.pl to be selective about disabled algorithms.
15172 Previously, it would create entries for disabled algorithms no
15177 * Added several new manual pages for SSL_* function.
15181 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15183 * In ssl23_get_client_hello, generate an error message when faced
15184 with an initial SSL 3.0/TLS record that is too small to contain the
15185 first two bytes of the ClientHello message, i.e. client_version.
15186 (Note that this is a pathologic case that probably has never happened
15187 in real life.) The previous approach was to use the version number
15188 from the record header as a substitute; but our protocol choice
15189 should not depend on that one because it is not authenticated
15190 by the Finished messages.
15194 * More robust randomness gathering functions for Windows.
15196 *Jeffrey Altman <jaltman@columbia.edu>*
15198 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15199 not set then we don't setup the error code for issuer check errors
15200 to avoid possibly overwriting other errors which the callback does
15201 handle. If an application does set the flag then we assume it knows
15202 what it is doing and can handle the new informational codes
15207 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15208 a general "ANY" type, as such it should be able to decode anything
15209 including tagged types. However it didn't check the class so it would
15210 wrongly interpret tagged types in the same way as their universal
15211 counterpart and unknown types were just rejected. Changed so that the
15212 tagged and unknown types are handled in the same way as a SEQUENCE:
15213 that is the encoding is stored intact. There is also a new type
15214 "V_ASN1_OTHER" which is used when the class is not universal, in this
15215 case we have no idea what the actual type is so we just lump them all
15220 * On VMS, stdout may very well lead to a file that is written to
15221 in a record-oriented fashion. That means that every write() will
15222 write a separate record, which will be read separately by the
15223 programs trying to read from it. This can be very confusing.
15225 The solution is to put a BIO filter in the way that will buffer
15226 text until a linefeed is reached, and then write everything a
15227 line at a time, so every record written will be an actual line,
15228 not chunks of lines and not (usually doesn't happen, but I've
15229 seen it once) several lines in one record. BIO_f_linebuffer() is
15232 Currently, it's a VMS-only method, because that's where it has
15233 been tested well enough.
15237 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15238 it can return incorrect results.
15239 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15240 but it was in 0.9.6-beta[12].)
15244 * Disable the check for content being present when verifying detached
15245 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15246 include zero length content when signing messages.
15250 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15251 BIO_ctrl (for BIO pairs).
15255 * Add DSO method for VMS.
15259 * Bug fix: Montgomery multiplication could produce results with the
15264 * Add RPM specification openssl.spec and modify it to build three
15265 packages. The default package contains applications, application
15266 documentation and run-time libraries. The devel package contains
15267 include files, static libraries and function documentation. The
15268 doc package contains the contents of the doc directory. The original
15269 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15273 * Add a large number of documentation files for many SSL routines.
15275 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15277 * Add a configuration entry for Sony News 4.
15279 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15281 * Don't set the two most significant bits to one when generating a
15282 random number < q in the DSA library.
15286 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15287 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15288 the underlying transport is blocking) if a handshake took place.
15289 (The default behaviour is needed by applications such as s_client
15290 and s_server that use select() to determine when to use SSL_read;
15291 but for applications that know in advance when to expect data, it
15292 just makes things more complicated.)
15296 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15301 * Add a few more EBCDIC conditionals that make `req` and `x509`
15302 work better on such systems.
15304 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15306 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15307 Update PKCS12_parse() so it copies the friendlyName and the
15308 keyid to the certificates aux info.
15312 * Fix bug in PKCS7_verify() which caused an infinite loop
15313 if there was more than one signature.
15315 *Sven Uszpelkat <su@celocom.de>*
15317 * Major change in util/mkdef.pl to include extra information
15318 about each symbol, as well as presenting variables as well
15319 as functions. This change means that there's n more need
15320 to rebuild the .num files when some algorithms are excluded.
15324 * Allow the verify time to be set by an application,
15325 rather than always using the current time.
15329 * Phase 2 verify code reorganisation. The certificate
15330 verify code now looks up an issuer certificate by a
15331 number of criteria: subject name, authority key id
15332 and key usage. It also verifies self signed certificates
15333 by the same criteria. The main comparison function is
15334 X509_check_issued() which performs these checks.
15336 Lot of changes were necessary in order to support this
15337 without completely rewriting the lookup code.
15339 Authority and subject key identifier are now cached.
15341 The LHASH 'certs' is X509_STORE has now been replaced
15342 by a STACK_OF(X509_OBJECT). This is mainly because an
15343 LHASH can't store or retrieve multiple objects with
15344 the same hash value.
15346 As a result various functions (which were all internal
15347 use only) have changed to handle the new X509_STORE
15348 structure. This will break anything that messed round
15349 with X509_STORE internally.
15351 The functions X509_STORE_add_cert() now checks for an
15352 exact match, rather than just subject name.
15354 The X509_STORE API doesn't directly support the retrieval
15355 of multiple certificates matching a given criteria, however
15356 this can be worked round by performing a lookup first
15357 (which will fill the cache with candidate certificates)
15358 and then examining the cache for matches. This is probably
15359 the best we can do without throwing out X509_LOOKUP
15360 entirely (maybe later...).
15362 The X509_VERIFY_CTX structure has been enhanced considerably.
15364 All certificate lookup operations now go via a get_issuer()
15365 callback. Although this currently uses an X509_STORE it
15366 can be replaced by custom lookups. This is a simple way
15367 to bypass the X509_STORE hackery necessary to make this
15368 work and makes it possible to use more efficient techniques
15369 in future. A very simple version which uses a simple
15370 STACK for its trusted certificate store is also provided
15371 using X509_STORE_CTX_trusted_stack().
15373 The verify_cb() and verify() callbacks now have equivalents
15374 in the X509_STORE_CTX structure.
15376 X509_STORE_CTX also has a 'flags' field which can be used
15377 to customise the verify behaviour.
15381 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15382 excludes S/MIME capabilities.
15386 * When a certificate request is read in keep a copy of the
15387 original encoding of the signed data and use it when outputting
15388 again. Signatures then use the original encoding rather than
15389 a decoded, encoded version which may cause problems if the
15390 request is improperly encoded.
15394 * For consistency with other BIO_puts implementations, call
15395 buffer_write(b, ...) directly in buffer_puts instead of calling
15398 In BIO_puts, increment b->num_write as in BIO_write.
15400 *Peter.Sylvester@EdelWeb.fr*
15402 * Fix BN_mul_word for the case where the word is 0. (We have to use
15403 BN_zero, we may not return a BIGNUM with an array consisting of
15404 words set to zero.)
15408 * Avoid calling abort() from within the library when problems are
15409 detected, except if preprocessor symbols have been defined
15410 (such as REF_CHECK, BN_DEBUG etc.).
15414 * New openssl application 'rsautl'. This utility can be
15415 used for low-level RSA operations. DER public key
15416 BIO/fp routines also added.
15420 * New Configure entry and patches for compiling on QNX 4.
15422 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15424 * A demo state-machine implementation was sponsored by
15425 Nuron (<http://www.nuron.com/>) and is now available in
15426 demos/state_machine.
15430 * New options added to the 'dgst' utility for signature
15431 generation and verification.
15435 * Unrecognized PKCS#7 content types are now handled via a
15436 catch all ASN1_TYPE structure. This allows unsupported
15437 types to be stored as a "blob" and an application can
15438 encode and decode it manually.
15442 * Fix various signed/unsigned issues to make a_strex.c
15443 compile under VC++.
15445 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15447 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15448 length if passed a buffer. ASN1_INTEGER_to_BN failed
15449 if passed a NULL BN and its argument was negative.
15451 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15453 * Modification to PKCS#7 encoding routines to output definite
15454 length encoding. Since currently the whole structures are in
15455 memory there's not real point in using indefinite length
15456 constructed encoding. However if OpenSSL is compiled with
15457 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15461 * Added BIO_vprintf() and BIO_vsnprintf().
15465 * Added more prefixes to parse for in the strings written
15466 through a logging bio, to cover all the levels that are available
15467 through syslog. The prefixes are now:
15469 PANIC, EMERG, EMR => LOG_EMERG
15470 ALERT, ALR => LOG_ALERT
15471 CRIT, CRI => LOG_CRIT
15472 ERROR, ERR => LOG_ERR
15473 WARNING, WARN, WAR => LOG_WARNING
15474 NOTICE, NOTE, NOT => LOG_NOTICE
15475 INFO, INF => LOG_INFO
15476 DEBUG, DBG => LOG_DEBUG
15478 and as before, if none of those prefixes are present at the
15479 beginning of the string, LOG_ERR is chosen.
15481 On Win32, the `LOG_*` levels are mapped according to this:
15483 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15484 LOG_WARNING => EVENTLOG_WARNING_TYPE
15485 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15489 * Made it possible to reconfigure with just the configuration
15490 argument "reconf" or "reconfigure". The command line arguments
15491 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15492 and are retrieved from there when reconfiguring.
15498 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15500 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15504 * The obj_dat.pl script was messing up the sorting of object
15505 names. The reason was that it compared the quoted version
15506 of strings as a result "OCSP" > "OCSP Signing" because
15507 " > SPACE. Changed script to store unquoted versions of
15508 names and add quotes on output. It was also omitting some
15509 names from the lookup table if they were given a default
15510 value (that is if SN is missing it is given the same
15511 value as LN and vice versa), these are now added on the
15512 grounds that if an object has a name we should be able to
15513 look it up. Finally added warning output when duplicate
15514 short or long names are found.
15518 * Changes needed for Tandem NSK.
15520 *Scott Uroff <scott@xypro.com>*
15522 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15523 RSA_padding_check_SSLv23(), special padding was never detected
15524 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15525 version rollback attacks was not effective.
15527 In s23_clnt.c, don't use special rollback-attack detection padding
15528 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15529 client; similarly, in s23_srvr.c, don't do the rollback check if
15530 SSL 2.0 is the only protocol enabled in the server.
15534 * Make it possible to get hexdumps of unprintable data with 'openssl
15535 asn1parse'. By implication, the functions ASN1_parse_dump() and
15536 BIO_dump_indent() are added.
15540 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15541 these print out strings and name structures based on various
15542 flags including RFC2253 support and proper handling of
15543 multibyte characters. Added options to the 'x509' utility
15544 to allow the various flags to be set.
15548 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15549 Also change the functions X509_cmp_current_time() and
15550 X509_gmtime_adj() work with an ASN1_TIME structure,
15551 this will enable certificates using GeneralizedTime in validity
15552 dates to be checked.
15556 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15557 negative public key encodings) on by default,
15558 NO_NEG_PUBKEY_BUG can be set to disable it.
15562 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15563 content octets. An i2c_ASN1_OBJECT is unnecessary because
15564 the encoding can be trivially obtained from the structure.
15568 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15569 not read locks (`CRYPTO_r_[un]lock`).
15573 * A first attempt at creating official support for shared
15574 libraries through configuration. I've kept it so the
15575 default is static libraries only, and the OpenSSL programs
15576 are always statically linked for now, but there are
15577 preparations for dynamic linking in place.
15578 This has been tested on Linux and Tru64.
15582 * Randomness polling function for Win9x, as described in:
15583 Peter Gutmann, Software Generation of Practically Strong
15588 * Fix so PRNG is seeded in req if using an already existing
15593 * New options to smime application. -inform and -outform
15594 allow alternative formats for the S/MIME message including
15595 PEM and DER. The -content option allows the content to be
15596 specified separately. This should allow things like Netscape
15597 form signing output easier to verify.
15601 * Fix the ASN1 encoding of tags using the 'long form'.
15605 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15606 STRING types. These convert content octets to and from the
15607 underlying type. The actual tag and length octets are
15608 already assumed to have been read in and checked. These
15609 are needed because all other string types have virtually
15610 identical handling apart from the tag. By having versions
15611 of the ASN1 functions that just operate on content octets
15612 IMPLICIT tagging can be handled properly. It also allows
15613 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15614 and ASN1_INTEGER are identical apart from the tag.
15618 * Change the handling of OID objects as follows:
15620 - New object identifiers are inserted in objects.txt, following
15621 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15622 - objects.pl is used to process obj_mac.num and create a new
15624 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15627 This is currently kind of a hack, and the perl code in objects.pl
15628 isn't very elegant, but it works as I intended. The simplest way
15629 to check that it worked correctly is to look in obj_dat.h and
15630 check the array nid_objs and make sure the objects haven't moved
15631 around (this is important!). Additions are OK, as well as
15632 consistent name changes.
15636 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15640 * Addition of the command line parameter '-rand file' to 'openssl req'.
15641 The given file adds to whatever has already been seeded into the
15642 random pool through the RANDFILE configuration file option or
15643 environment variable, or the default random state file.
15647 * mkstack.pl now sorts each macro group into lexical order.
15648 Previously the output order depended on the order the files
15649 appeared in the directory, resulting in needless rewriting
15654 * Patches to make OpenSSL compile under Win32 again. Mostly
15655 work arounds for the VC++ problem that it treats func() as
15656 func(void). Also stripped out the parts of mkdef.pl that
15657 added extra typesafe functions: these no longer exist.
15661 * Reorganisation of the stack code. The macros are now all
15662 collected in safestack.h . Each macro is defined in terms of
15663 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15664 DEBUG_SAFESTACK is now handled in terms of function casts,
15665 this has the advantage of retaining type safety without the
15666 use of additional functions. If DEBUG_SAFESTACK is not defined
15667 then the non typesafe macros are used instead. Also modified the
15668 mkstack.pl script to handle the new form. Needs testing to see
15669 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15670 the default if no major problems. Similar behaviour for ASN1_SET_OF
15671 and PKCS12_STACK_OF.
15675 * When some versions of IIS use the 'NET' form of private key the
15676 key derivation algorithm is different. Normally MD5(password) is
15677 used as a 128 bit RC4 key. In the modified case
15678 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15679 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15680 as the old Netscape_RSA functions except they have an additional
15681 'sgckey' parameter which uses the modified algorithm. Also added
15682 an -sgckey command line option to the rsa utility. Thanks to
15683 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15684 algorithm to openssl-dev.
15688 * The evp_local.h macros were using 'c.##kname' which resulted in
15689 invalid expansion on some systems (SCO 5.0.5 for example).
15690 Corrected to 'c.kname'.
15692 *Phillip Porch <root@theporch.com>*
15694 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15695 a STACK of email addresses from a certificate or request, these look
15696 in the subject name and the subject alternative name extensions and
15697 omit any duplicate addresses.
15701 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15702 This makes DSA verification about 2 % faster.
15706 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15707 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15708 plus overhead for 1024 bit moduli).
15709 This makes exponentiations about 0.5 % faster for 1024 bit
15710 exponents (as measured by "openssl speed rsa2048").
15714 * Rename memory handling macros to avoid conflicts with other
15716 Malloc => OPENSSL_malloc
15717 Malloc_locked => OPENSSL_malloc_locked
15718 Realloc => OPENSSL_realloc
15719 Free => OPENSSL_free
15723 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15724 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15728 * CygWin32 support.
15730 *John Jarvie <jjarvie@newsguy.com>*
15732 * The type-safe stack code has been rejigged. It is now only compiled
15733 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15734 by default all type-specific stack functions are "#define"d back to
15735 standard stack functions. This results in more streamlined output
15736 but retains the type-safety checking possibilities of the original
15741 * The STACK code has been cleaned up, and certain type declarations
15742 that didn't make a lot of sense have been brought in line. This has
15743 also involved a cleanup of sorts in safestack.h to more correctly
15744 map type-safe stack functions onto their plain stack counterparts.
15745 This work has also resulted in a variety of "const"ifications of
15746 lots of the code, especially `_cmp` operations which should normally
15747 be prototyped with "const" parameters anyway.
15751 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15752 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15753 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15754 where all of 'md' is used each time the PRNG is used, but 'state'
15755 is used only indexed by a cyclic counter. As entropy may not be
15756 well distributed from the beginning, 'md' is important as a
15757 chaining variable. However, the output function chains only half
15758 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15759 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15760 in all of 'state' being rewritten, with the new values depending
15761 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15765 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15766 the handshake is continued after ssl_verify_cert_chain();
15767 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15768 can lead to 'unexplainable' connection aborts later.
15770 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15772 * Major EVP API cipher revision.
15773 Add hooks for extra EVP features. This allows various cipher
15774 parameters to be set in the EVP interface. Support added for variable
15775 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15776 setting of RC2 and RC5 parameters.
15778 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15781 Remove lots of duplicated code from the EVP library. For example *every*
15782 cipher init() function handles the 'iv' in the same way according to the
15783 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15784 for CFB and OFB modes they zero ctx->num.
15786 New functionality allows removal of S/MIME code RC2 hack.
15788 Most of the routines have the same form and so can be declared in terms
15791 By shifting this to the top level EVP_CipherInit() it can be removed from
15792 all individual ciphers. If the cipher wants to handle IVs or keys
15793 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15796 Change lots of functions like EVP_EncryptUpdate() to now return a
15797 value: although software versions of the algorithms cannot fail
15798 any installed hardware versions can.
15802 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15803 this option is set, tolerate broken clients that send the negotiated
15804 protocol version number instead of the requested protocol version
15809 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15810 i.e. non-zero for export ciphersuites, zero otherwise.
15811 Previous versions had this flag inverted, inconsistent with
15812 rsa_tmp_cb (..._TMP_RSA_CB).
15814 *Bodo Moeller; problem reported by Amit Chopra*
15816 * Add missing DSA library text string. Work around for some IIS
15817 key files with invalid SEQUENCE encoding.
15821 * Add a document (doc/standards.txt) that list all kinds of standards
15822 and so on that are implemented in OpenSSL.
15826 * Enhance c_rehash script. Old version would mishandle certificates
15827 with the same subject name hash and wouldn't handle CRLs at all.
15828 Added -fingerprint option to crl utility, to support new c_rehash
15833 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15837 * Fix for SSL server purpose checking. Server checking was
15838 rejecting certificates which had extended key usage present
15839 but no ssl client purpose.
15841 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15843 * Make PKCS#12 code work with no password. The PKCS#12 spec
15844 is a little unclear about how a blank password is handled.
15845 Since the password in encoded as a BMPString with terminating
15846 double NULL a zero length password would end up as just the
15847 double NULL. However no password at all is different and is
15848 handled differently in the PKCS#12 key generation code. NS
15849 treats a blank password as zero length. MSIE treats it as no
15850 password on export: but it will try both on import. We now do
15851 the same: PKCS12_parse() tries zero length and no password if
15852 the password is set to "" or NULL (NULL is now a valid password:
15853 it wasn't before) as does the pkcs12 application.
15857 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15858 perror when PEM_read_bio_X509_REQ fails, the error message must
15859 be obtained from the error queue.
15863 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15864 it in ERR_remove_state if appropriate, and change ERR_get_state
15865 accordingly to avoid race conditions (this is necessary because
15866 thread_hash is no longer constant once set).
15870 * Bugfix for linux-elf makefile.one.
15874 * RSA_get_default_method() will now cause a default
15875 RSA_METHOD to be chosen if one doesn't exist already.
15876 Previously this was only set during a call to RSA_new()
15877 or RSA_new_method(NULL) meaning it was possible for
15878 RSA_get_default_method() to return NULL.
15882 * Added native name translation to the existing DSO code
15883 that will convert (if the flag to do so is set) filenames
15884 that are sufficiently small and have no path information
15885 into a canonical native form. Eg. "blah" converted to
15886 "libblah.so" or "blah.dll" etc.
15890 * New function ERR_error_string_n(e, buf, len) which is like
15891 ERR_error_string(e, buf), but writes at most 'len' bytes
15892 including the 0 terminator. For ERR_error_string_n, 'buf'
15895 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15897 * CONF library reworked to become more general. A new CONF
15898 configuration file reader "class" is implemented as well as a
15899 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15900 old `CONF_*` functions are still there, but are reimplemented to
15901 work in terms of the new functions. Also, a set of functions
15902 to handle the internal storage of the configuration data is
15903 provided to make it easier to write new configuration file
15904 reader "classes" (I can definitely see something reading a
15905 configuration file in XML format, for example), called `_CONF_*`,
15906 or "the configuration storage API"...
15908 The new configuration file reading functions are:
15910 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15911 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15913 NCONF_default, NCONF_WIN32
15915 NCONF_dump_fp, NCONF_dump_bio
15917 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15918 NCONF_new creates a new CONF object. This works in the same way
15919 as other interfaces in OpenSSL, like the BIO interface.
15920 `NCONF_dump_*` dump the internal storage of the configuration file,
15921 which is useful for debugging. All other functions take the same
15922 arguments as the old `CONF_*` functions with the exception of the
15923 first that must be a `CONF *` instead of a `LHASH *`.
15925 To make it easier to use the new classes with the old `CONF_*` functions,
15926 the function CONF_set_default_method is provided.
15930 * Add '-tls1' option to 'openssl ciphers', which was already
15931 mentioned in the documentation but had not been implemented.
15932 (This option is not yet really useful because even the additional
15933 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15937 * Initial DSO code added into libcrypto for letting OpenSSL (and
15938 OpenSSL-based applications) load shared libraries and bind to
15939 them in a portable way.
15941 *Geoff Thorpe, with contributions from Richard Levitte*
15943 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15945 * Make sure _lrotl and _lrotr are only used with MSVC.
15947 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15948 (the default implementation of RAND_status).
15950 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15951 to '-clrext' (= clear extensions), as intended and documented.
15952 *Bodo Moeller; inconsistency pointed out by Michael Attili
15953 <attili@amaxo.com>*
15955 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15956 was larger than the MD block size.
15958 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15960 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15961 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15962 using the passed key: if the passed key was a private key the result
15963 of X509_print(), for example, would be to print out all the private key
15968 * des_quad_cksum() byte order bug fix.
15969 *Ulf Möller, using the problem description in krb4-0.9.7, where
15970 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15972 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15975 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15977 * For easily testing in shell scripts whether some command
15978 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15979 returns with exit code 0 iff no command of the given name is available.
15980 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15981 the output goes to stdout and nothing is printed to stderr.
15982 Additional arguments are always ignored.
15984 Since for each cipher there is a command of the same name,
15985 the 'no-cipher' compilation switches can be tested this way.
15987 ('openssl no-XXX' is not able to detect pseudo-commands such
15988 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15992 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15996 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15997 is set; it will be thrown away anyway because each handshake creates
15999 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16000 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16001 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16002 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16006 * New s_client option -ign_eof: EOF at stdin is ignored, and
16007 'Q' and 'R' lose their special meanings (quit/renegotiate).
16008 This is part of what -quiet does; unlike -quiet, -ign_eof
16009 does not suppress any output.
16013 * Add compatibility options to the purpose and trust code. The
16014 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16015 accepts a certificate or CA, this was the previous behaviour,
16016 with all the associated security issues.
16018 X509_TRUST_COMPAT is the old trust behaviour: only and
16019 automatically trust self signed roots in certificate store. A
16020 new trust setting X509_TRUST_DEFAULT is used to specify that
16021 a purpose has no associated trust setting and it should instead
16022 use the value in the default purpose.
16026 * Fix the PKCS#8 DSA private key code so it decodes keys again
16027 and fix a memory leak.
16031 * In util/mkerr.pl (which implements 'make errors'), preserve
16032 reason strings from the previous version of the .c file, as
16033 the default to have only downcase letters (and digits) in
16034 automatically generated reasons codes is not always appropriate.
16038 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16039 using strerror. Previously, ERR_reason_error_string() returned
16040 library names as reason strings for SYSerr; but SYSerr is a special
16041 case where small numbers are errno values, not library numbers.
16045 * Add '-dsaparam' option to 'openssl dhparam' application. This
16046 converts DSA parameters into DH parameters. (When creating parameters,
16047 DSA_generate_parameters is used.)
16051 * Include 'length' (recommended exponent length) in C code generated
16052 by 'openssl dhparam -C'.
16056 * The second argument to set_label in perlasm was already being used
16057 so couldn't be used as a "file scope" flag. Moved to third argument
16062 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16063 instead of RAND_bytes for encryption IVs and salts.
16067 * Include RAND_status() into RAND_METHOD instead of implementing
16068 it only for md_rand.c Otherwise replacing the PRNG by calling
16069 RAND_set_rand_method would be impossible.
16073 * Don't let DSA_generate_key() enter an infinite loop if the random
16074 number generation fails.
16078 * New 'rand' application for creating pseudo-random output.
16082 * Added configuration support for Linux/IA64
16084 *Rolf Haberrecker <rolf@suse.de>*
16086 * Assembler module support for Mingw32.
16090 * Shared library support for HPUX (in shlib/).
16092 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16094 * Shared library support for Solaris gcc.
16096 *Lutz Behnke <behnke@trustcenter.de>*
16098 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16100 * PKCS7_encrypt() was adding text MIME headers twice because they
16101 were added manually and by SMIME_crlf_copy().
16105 * In bntest.c don't call BN_rand with zero bits argument.
16107 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16109 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16110 case was implemented. This caused BN_div_recp() to fail occasionally.
16114 * Add an optional second argument to the set_label() in the perl
16115 assembly language builder. If this argument exists and is set
16116 to 1 it signals that the assembler should use a symbol whose
16117 scope is the entire file, not just the current function. This
16118 is needed with MASM which uses the format label:: for this scope.
16120 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16122 * Change the ASN1 types so they are typedefs by default. Before
16123 almost all types were #define'd to ASN1_STRING which was causing
16124 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16129 * Change names of new functions to the new get1/get0 naming
16130 convention: After 'get1', the caller owns a reference count
16131 and has to call `..._free`; 'get0' returns a pointer to some
16132 data structure without incrementing reference counters.
16133 (Some of the existing 'get' functions increment a reference
16134 counter, some don't.)
16135 Similarly, 'set1' and 'add1' functions increase reference
16136 counters or duplicate objects.
16140 * Allow for the possibility of temp RSA key generation failure:
16141 the code used to assume it always worked and crashed on failure.
16145 * Fix potential buffer overrun problem in BIO_printf().
16146 *Ulf Möller, using public domain code by Patrick Powell; problem
16147 pointed out by David Sacerdote <das33@cornell.edu>*
16149 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16150 RAND_egd() and RAND_status(). In the command line application,
16151 the EGD socket can be specified like a seed file using RANDFILE
16156 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16157 Some CAs (e.g. Verisign) distribute certificates in this form.
16161 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16162 list to exclude them. This means that no special compilation option
16163 is needed to use anonymous DH: it just needs to be included in the
16168 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16169 EVP_MD_type. The old functionality is available in a new macro called
16170 EVP_MD_md(). Change code that uses it and update docs.
16174 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16175 where the `void *` argument is replaced by a function pointer argument.
16176 Previously `void *` was abused to point to functions, which works on
16177 many platforms, but is not correct. As these functions are usually
16178 called by macros defined in OpenSSL header files, most source code
16179 should work without changes.
16183 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16184 sections with information on -D... compiler switches used for
16185 compiling the library so that applications can see them. To enable
16186 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16187 must be defined. E.g.,
16188 #define OPENSSL_ALGORITHM_DEFINES
16189 #include <openssl/opensslconf.h>
16190 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16192 *Richard Levitte, Ulf and Bodo Möller*
16194 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16199 * Change the 'other' type in certificate aux info to a STACK_OF
16200 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16201 the required ASN1 format: arbitrary types determined by an OID.
16205 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16206 argument to 'req'. This is not because the function is newer or
16207 better than others it just uses the work 'NEW' in the certificate
16208 request header lines. Some software needs this.
16212 * Reorganise password command line arguments: now passwords can be
16213 obtained from various sources. Delete the PEM_cb function and make
16214 it the default behaviour: i.e. if the callback is NULL and the
16215 usrdata argument is not NULL interpret it as a null terminated pass
16216 phrase. If usrdata and the callback are NULL then the pass phrase
16217 is prompted for as usual.
16221 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16222 the support is automatically enabled. The resulting binaries will
16223 autodetect the card and use it if present.
16225 *Ben Laurie and Compaq Inc.*
16227 * Work around for Netscape hang bug. This sends certificate request
16228 and server done in one record. Since this is perfectly legal in the
16229 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16230 the bugs/SSLv3 entry for more info.
16234 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16238 * Add -rand argument to smime and pkcs12 applications and read/write
16243 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16247 * Add command line password options to the remaining applications.
16251 * Bug fix for BN_div_recp() for numerators with an even number of
16256 * More tests in bntest.c, and changed test_bn output.
16260 * ./config recognizes MacOS X now.
16264 * Bug fix for BN_div() when the first words of num and divisor are
16265 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16269 * Add support for various broken PKCS#8 formats, and command line
16270 options to produce them.
16274 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16275 get temporary BIGNUMs from a BN_CTX.
16279 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16284 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16285 include a #define from the old name to the new. The original intent
16286 was that statically linked binaries could for example just call
16287 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16288 link with digests. This never worked because SSLeay_add_all_digests()
16289 and SSLeay_add_all_ciphers() were in the same source file so calling
16290 one would link with the other. They are now in separate source files.
16294 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16298 * Use a less unusual form of the Miller-Rabin primality test (it used
16299 a binary algorithm for exponentiation integrated into the Miller-Rabin
16300 loop, our standard modexp algorithms are faster).
16304 * Support for the EBCDIC character set completed.
16306 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16308 * Source code cleanups: use const where appropriate, eliminate casts,
16309 use `void *` instead of `char *` in lhash.
16313 * Bugfix: ssl3_send_server_key_exchange was not restartable
16314 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16315 this the server could overwrite ephemeral keys that the client
16320 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16321 using 50 iterations of the Rabin-Miller test.
16323 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16324 iterations of the Rabin-Miller test as required by the appendix
16325 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16326 As BN_is_prime_fasttest includes trial division, DSA parameter
16327 generation becomes much faster.
16329 This implies a change for the callback functions in DSA_is_prime
16330 and DSA_generate_parameters: The callback function is called once
16331 for each positive witness in the Rabin-Miller test, not just
16332 occasionally in the inner loop; and the parameters to the
16333 callback function now provide an iteration count for the outer
16334 loop rather than for the current invocation of the inner loop.
16335 DSA_generate_parameters additionally can call the callback
16336 function with an 'iteration count' of -1, meaning that a
16337 candidate has passed the trial division test (when q is generated
16338 from an application-provided seed, trial division is skipped).
16342 * New function BN_is_prime_fasttest that optionally does trial
16343 division before starting the Rabin-Miller test and has
16344 an additional BN_CTX * argument (whereas BN_is_prime always
16345 has to allocate at least one BN_CTX).
16346 'callback(1, -1, cb_arg)' is called when a number has passed the
16347 trial division stage.
16351 * Fix for bug in CRL encoding. The validity dates weren't being handled
16356 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16360 * New function BN_pseudo_rand().
16364 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16365 bignum version of BN_from_montgomery() with the working code from
16366 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16371 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16372 made it impossible to use the same SSL_SESSION data structure in
16373 SSL2 clients in multiple threads.
16377 * The return value of RAND_load_file() no longer counts bytes obtained
16378 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16379 to seed the PRNG (previously an explicit byte count was required).
16381 *Ulf Möller, Bodo Möller*
16383 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16384 used `char *` instead of `void *` and had casts all over the place.
16388 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16392 * Retain source code compatibility for BN_prime_checks macro:
16393 BN_is_prime(..., BN_prime_checks, ...) now uses
16394 BN_prime_checks_for_size to determine the appropriate number of
16395 Rabin-Miller iterations.
16399 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16400 DH_CHECK_P_NOT_SAFE_PRIME.
16401 (Check if this is true? OpenPGP calls them "strong".)
16405 * Merge the functionality of "dh" and "gendh" programs into a new program
16406 "dhparam". The old programs are retained for now but will handle DH keys
16407 (instead of parameters) in future.
16411 * Make the ciphers, s_server and s_client programs check the return values
16412 when a new cipher list is set.
16416 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16417 ciphers. Before when the 56bit ciphers were enabled the sorting was
16420 The syntax for the cipher sorting has been extended to support sorting by
16421 cipher-strength (using the strength_bits hard coded in the tables).
16422 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16424 Fix a bug in the cipher-command parser: when supplying a cipher command
16425 string with an "undefined" symbol (neither command nor alphanumeric
16426 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16427 an error is flagged.
16429 Due to the strength-sorting extension, the code of the
16430 ssl_create_cipher_list() function was completely rearranged. I hope that
16431 the readability was also increased :-)
16433 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16435 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16436 for the first serial number and places 2 in the serial number file. This
16437 avoids problems when the root CA is created with serial number zero and
16438 the first user certificate has the same issuer name and serial number
16443 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16444 the new code. Add documentation for this stuff.
16448 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16449 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16450 structures and behave in an analogous way to the X509v3 functions:
16451 they shouldn't be called directly but wrapper functions should be used
16454 So we also now have some wrapper functions that call the X509at functions
16455 when passed certificate requests. (TO DO: similar things can be done with
16456 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16457 things. Some of these need some d2i or i2d and print functionality
16458 because they handle more complex structures.)
16462 * Add missing #ifndefs that caused missing symbols when building libssl
16463 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16464 NO_RSA in `ssl/s2*.c`.
16466 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16468 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16469 has a return value which indicates the quality of the random data
16470 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16471 error queue. New function RAND_pseudo_bytes() generates output that is
16472 guaranteed to be unique but not unpredictable. RAND_add is like
16473 RAND_seed, but takes an extra argument for an entropy estimate
16474 (RAND_seed always assumes full entropy).
16478 * Do more iterations of Rabin-Miller probable prime test (specifically,
16479 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16480 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16481 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16482 false-positive rate of at most 2^-80 for random input.
16486 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16490 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16491 in the 0.9.5 release), this returns the chain
16492 from an X509_CTX structure with a dup of the stack and all
16493 the X509 reference counts upped: so the stack will exist
16494 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16497 Also make SSL_SESSION_print() print out the verify return
16502 * Add manpage for the pkcs12 command. Also change the default
16503 behaviour so MAC iteration counts are used unless the new
16504 -nomaciter option is used. This improves file security and
16505 only older versions of MSIE (4.0 for example) need it.
16509 * Honor the no-xxx Configure options when creating .DEF files.
16513 * Add PKCS#10 attributes to field table: challengePassword,
16514 unstructuredName and unstructuredAddress. These are taken from
16515 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16516 international characters are used.
16518 More changes to X509_ATTRIBUTE code: allow the setting of types
16519 based on strings. Remove the 'loc' parameter when adding
16520 attributes because these will be a SET OF encoding which is sorted
16525 * Initial changes to the 'req' utility to allow request generation
16526 automation. This will allow an application to just generate a template
16527 file containing all the field values and have req construct the
16530 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16531 used all over the place including certificate requests and PKCS#7
16532 structures. They are currently handled manually where necessary with
16533 some primitive wrappers for PKCS#7. The new functions behave in a
16534 manner analogous to the X509 extension functions: they allow
16535 attributes to be looked up by NID and added.
16537 Later something similar to the X509V3 code would be desirable to
16538 automatically handle the encoding, decoding and printing of the
16539 more complex types. The string types like challengePassword can
16540 be handled by the string table functions.
16542 Also modified the multi byte string table handling. Now there is
16543 a 'global mask' which masks out certain types. The table itself
16544 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16545 is useful when for example there is only one permissible type
16546 (as in countryName) and using the mask might result in no valid
16551 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16552 SSL_get_peer_finished to allow applications to obtain the latest
16553 Finished messages sent to the peer or expected from the peer,
16554 respectively. (SSL_get_peer_finished is usually the Finished message
16555 actually received from the peer, otherwise the protocol will be aborted.)
16557 As the Finished message are message digests of the complete handshake
16558 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16559 be used for external authentication procedures when the authentication
16560 provided by SSL/TLS is not desired or is not enough.
16564 * Enhanced support for Alpha Linux is added. Now ./config checks if
16565 the host supports BWX extension and if Compaq C is present on the
16566 $PATH. Just exploiting of the BWX extension results in 20-30%
16567 performance kick for some algorithms, e.g. DES and RC4 to mention
16568 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16573 * Add support for MS "fast SGC". This is arguably a violation of the
16574 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16575 weak crypto and after checking the certificate is SGC a second one
16576 with strong crypto. MS SGC stops the first handshake after receiving
16577 the server certificate message and sends a second client hello. Since
16578 a server will typically do all the time consuming operations before
16579 expecting any further messages from the client (server key exchange
16580 is the most expensive) there is little difference between the two.
16582 To get OpenSSL to support MS SGC we have to permit a second client
16583 hello message after we have sent server done. In addition we have to
16584 reset the MAC if we do get this second client hello.
16588 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16589 if a DER encoded private key is RSA or DSA traditional format. Changed
16590 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16591 format DER encoded private key. Newer code should use PKCS#8 format which
16592 has the key type encoded in the ASN1 structure. Added DER private key
16593 support to pkcs8 application.
16597 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16598 ciphersuites has been selected (as required by the SSL 3/TLS 1
16599 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16600 is set, we interpret this as a request to violate the specification
16601 (the worst that can happen is a handshake failure, and 'correct'
16602 behaviour would result in a handshake failure anyway).
16606 * In SSL_CTX_add_session, take into account that there might be multiple
16607 SSL_SESSION structures with the same session ID (e.g. when two threads
16608 concurrently obtain them from an external cache).
16609 The internal cache can handle only one SSL_SESSION with a given ID,
16610 so if there's a conflict, we now throw out the old one to achieve
16615 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16616 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16617 some routines that use cipher OIDs: some ciphers do not have OIDs
16618 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16623 * Simplify the trust setting structure and code. Now we just have
16624 two sequences of OIDs for trusted and rejected settings. These will
16625 typically have values the same as the extended key usage extension
16626 and any application specific purposes.
16628 The trust checking code now has a default behaviour: it will just
16629 check for an object with the same NID as the passed id. Functions can
16630 be provided to override either the default behaviour or the behaviour
16631 for a given id. SSL client, server and email already have functions
16632 in place for compatibility: they check the NID and also return "trusted"
16633 if the certificate is self signed.
16637 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16638 traditional format into an EVP_PKEY structure.
16642 * Add a password callback function PEM_cb() which either prompts for
16643 a password if usr_data is NULL or otherwise assumes it is a null
16644 terminated password. Allow passwords to be passed on command line
16645 environment or config files in a few more utilities.
16649 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16650 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16651 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16652 Update documentation.
16656 * Support for ASN1 "NULL" type. This could be handled before by using
16657 ASN1_TYPE but there wasn't any function that would try to read a NULL
16658 and produce an error if it couldn't. For compatibility we also have
16659 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16660 don't allocate anything because they don't need to.
16664 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16667 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16669 * Rebuild of the memory allocation routines used by OpenSSL code and
16670 possibly others as well. The purpose is to make an interface that
16671 provide hooks so anyone can build a separate set of allocation and
16672 deallocation routines to be used by OpenSSL, for example memory
16673 pool implementations, or something else, which was previously hard
16674 since Malloc(), Realloc() and Free() were defined as macros having
16675 the values malloc, realloc and free, respectively (except for Win32
16676 compilations). The same is provided for memory debugging code.
16677 OpenSSL already comes with functionality to find memory leaks, but
16678 this gives people a chance to debug other memory problems.
16680 With these changes, a new set of functions and macros have appeared:
16682 CRYPTO_set_mem_debug_functions() [F]
16683 CRYPTO_get_mem_debug_functions() [F]
16684 CRYPTO_dbg_set_options() [F]
16685 CRYPTO_dbg_get_options() [F]
16686 CRYPTO_malloc_debug_init() [M]
16688 The memory debug functions are NULL by default, unless the library
16689 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16690 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16691 gives the standard debugging functions that come with OpenSSL) or
16692 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16693 provided by the library user) must be used. When the standard
16694 debugging functions are used, CRYPTO_dbg_set_options can be used to
16695 request additional information:
16696 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16697 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16699 Also, things like CRYPTO_set_mem_functions will always give the
16700 expected result (the new set of functions is used for allocation
16701 and deallocation) at all times, regardless of platform and compiler
16704 To finish it up, some functions that were never use in any other
16705 way than through macros have a new API and new semantic:
16707 CRYPTO_dbg_malloc()
16708 CRYPTO_dbg_realloc()
16711 All macros of value have retained their old syntax.
16713 *Richard Levitte and Bodo Moeller*
16715 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16716 ordering of SMIMECapabilities wasn't in "strength order" and there
16717 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16722 * Some ASN1 types with illegal zero length encoding (INTEGER,
16723 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16725 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16727 * Merge in my S/MIME library for OpenSSL. This provides a simple
16728 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16729 functionality to handle multipart/signed properly) and a utility
16730 called 'smime' to call all this stuff. This is based on code I
16731 originally wrote for Celo who have kindly allowed it to be
16732 included in OpenSSL.
16736 * Add variants des_set_key_checked and des_set_key_unchecked of
16737 des_set_key (aka des_key_sched). Global variable des_check_key
16738 decides which of these is called by des_set_key; this way
16739 des_check_key behaves as it always did, but applications and
16740 the library itself, which was buggy for des_check_key == 1,
16741 have a cleaner way to pick the version they need.
16745 * New function PKCS12_newpass() which changes the password of a
16750 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16751 dynamic mix. In both cases the ids can be used as an index into the
16752 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16753 functions so they accept a list of the field values and the
16754 application doesn't need to directly manipulate the X509_TRUST
16759 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16764 * Modify the way the V3 extension code looks up extensions. This now
16765 works in a similar way to the object code: we have some "standard"
16766 extensions in a static table which is searched with OBJ_bsearch()
16767 and the application can add dynamic ones if needed. The file
16768 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16769 updated whenever a new extension is added to the core code and kept
16770 in ext_nid order. There is a simple program 'tabtest.c' which checks
16771 this. New extensions are not added too often so this file can readily
16772 be maintained manually.
16774 There are two big advantages in doing things this way. The extensions
16775 can be looked up immediately and no longer need to be "added" using
16776 X509V3_add_standard_extensions(): this function now does nothing.
16777 Side note: I get *lots* of email saying the extension code doesn't
16778 work because people forget to call this function.
16779 Also no dynamic allocation is done unless new extensions are added:
16780 so if we don't add custom extensions there is no need to call
16781 X509V3_EXT_cleanup().
16785 * Modify enc utility's salting as follows: make salting the default. Add a
16786 magic header, so unsalted files fail gracefully instead of just decrypting
16787 to garbage. This is because not salting is a big security hole, so people
16788 should be discouraged from doing it.
16792 * Fixes and enhancements to the 'x509' utility. It allowed a message
16793 digest to be passed on the command line but it only used this
16794 parameter when signing a certificate. Modified so all relevant
16795 operations are affected by the digest parameter including the
16796 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16797 DSA key was used because it didn't fix the digest.
16801 * Initial certificate chain verify code. Currently tests the untrusted
16802 certificates for consistency with the verify purpose (which is set
16803 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16805 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16806 this is because it will reject chains with invalid extensions whereas
16807 every previous version of OpenSSL and SSLeay made no checks at all.
16809 Trust code: checks the root CA for the relevant trust settings. Trust
16810 settings have an initial value consistent with the verify purpose: e.g.
16811 if the verify purpose is for SSL client use it expects the CA to be
16812 trusted for SSL client use. However the default value can be changed to
16813 permit custom trust settings: one example of this would be to only trust
16814 certificates from a specific "secure" set of CAs.
16816 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16817 which should be used for version portability: especially since the
16818 verify structure is likely to change more often now.
16820 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16821 to set them. If not set then assume SSL clients will verify SSL servers
16824 Two new options to the verify program: -untrusted allows a set of
16825 untrusted certificates to be passed in and -purpose which sets the
16826 intended purpose of the certificate. If a purpose is set then the
16827 new chain verify code is used to check extension consistency.
16831 * Support for the authority information access extension.
16835 * Modify RSA and DSA PEM read routines to transparently handle
16836 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16837 public keys in a format compatible with certificate
16838 SubjectPublicKeyInfo structures. Unfortunately there were already
16839 functions called *_PublicKey_* which used various odd formats so
16840 these are retained for compatibility: however the DSA variants were
16841 never in a public release so they have been deleted. Changed dsa/rsa
16842 utilities to handle the new format: note no releases ever handled public
16843 keys so we should be OK.
16845 The primary motivation for this change is to avoid the same fiasco
16846 that dogs private keys: there are several incompatible private key
16847 formats some of which are standard and some OpenSSL specific and
16848 require various evil hacks to allow partial transparent handling and
16849 even then it doesn't work with DER formats. Given the option anything
16850 other than PKCS#8 should be dumped: but the other formats have to
16851 stay in the name of compatibility.
16853 With public keys and the benefit of hindsight one standard format
16854 is used which works with EVP_PKEY, RSA or DSA structures: though
16855 it clearly returns an error if you try to read the wrong kind of key.
16857 Added a -pubkey option to the 'x509' utility to output the public key.
16858 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16859 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16860 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16861 that do the same as the `EVP_PKEY_assign_*()` except they up the
16862 reference count of the added key (they don't "swallow" the
16867 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16868 CRLs would fail if the file contained no certificates or no CRLs:
16869 added a new function to read in both types and return the number
16870 read: this means that if none are read it will be an error. The
16871 DER versions of the certificate and CRL reader would always fail
16872 because it isn't possible to mix certificates and CRLs in DER format
16873 without choking one or the other routine. Changed this to just read
16874 a certificate: this is the best we can do. Also modified the code
16875 in `apps/verify.c` to take notice of return codes: it was previously
16876 attempting to read in certificates from NULL pointers and ignoring
16877 any errors: this is one reason why the cert and CRL reader seemed
16878 to work. It doesn't check return codes from the default certificate
16879 routines: these may well fail if the certificates aren't installed.
16883 * Code to support otherName option in GeneralName.
16887 * First update to verify code. Change the verify utility
16888 so it warns if it is passed a self signed certificate:
16889 for consistency with the normal behaviour. X509_verify
16890 has been modified to it will now verify a self signed
16891 certificate if *exactly* the same certificate appears
16892 in the store: it was previously impossible to trust a
16893 single self signed certificate. This means that:
16894 openssl verify ss.pem
16895 now gives a warning about a self signed certificate but
16896 openssl verify -CAfile ss.pem ss.pem
16901 * For servers, store verify_result in SSL_SESSION data structure
16902 (and add it to external session representation).
16903 This is needed when client certificate verifications fails,
16904 but an application-provided verification callback (set by
16905 SSL_CTX_set_cert_verify_callback) allows accepting the session
16906 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16907 but returns 1): When the session is reused, we have to set
16908 ssl->verify_result to the appropriate error code to avoid
16911 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16913 * Fix a bug in the new PKCS#7 code: it didn't consider the
16914 case in PKCS7_dataInit() where the signed PKCS7 structure
16915 didn't contain any existing data because it was being created.
16917 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16919 * Add a salt to the key derivation routines in enc.c. This
16920 forms the first 8 bytes of the encrypted file. Also add a
16921 -S option to allow a salt to be input on the command line.
16925 * New function X509_cmp(). Oddly enough there wasn't a function
16926 to compare two certificates. We do this by working out the SHA1
16927 hash and comparing that. X509_cmp() will be needed by the trust
16932 * SSL_get1_session() is like SSL_get_session(), but increments
16933 the reference count in the SSL_SESSION returned.
16935 *Geoff Thorpe <geoff@eu.c2.net>*
16937 * Fix for 'req': it was adding a null to request attributes.
16938 Also change the X509_LOOKUP and X509_INFO code to handle
16939 certificate auxiliary information.
16943 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16948 * Add the possibility to add extra information to the memory leak
16949 detecting output, to form tracebacks, showing from where each
16950 allocation was originated: CRYPTO_push_info("constant string") adds
16951 the string plus current file name and line number to a per-thread
16952 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16953 is like calling CYRPTO_pop_info() until the stack is empty.
16954 Also updated memory leak detection code to be multi-thread-safe.
16958 * Add options -text and -noout to pkcs7 utility and delete the
16959 encryption options which never did anything. Update docs.
16963 * Add options to some of the utilities to allow the pass phrase
16964 to be included on either the command line (not recommended on
16965 OSes like Unix) or read from the environment. Update the
16966 manpages and fix a few bugs.
16970 * Add a few manpages for some of the openssl commands.
16974 * Fix the -revoke option in ca. It was freeing up memory twice,
16975 leaking and not finding already revoked certificates.
16979 * Extensive changes to support certificate auxiliary information.
16980 This involves the use of X509_CERT_AUX structure and X509_AUX
16981 functions. An X509_AUX function such as PEM_read_X509_AUX()
16982 can still read in a certificate file in the usual way but it
16983 will also read in any additional "auxiliary information". By
16984 doing things this way a fair degree of compatibility can be
16985 retained: existing certificates can have this information added
16986 using the new 'x509' options.
16988 Current auxiliary information includes an "alias" and some trust
16989 settings. The trust settings will ultimately be used in enhanced
16990 certificate chain verification routines: currently a certificate
16991 can only be trusted if it is self signed and then it is trusted
16996 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16997 The problem was that one of the replacement routines had not been working
16998 since SSLeay releases. For now the offending routine has been replaced
16999 with non-optimised assembler. Even so, this now gives around 95%
17000 performance improvement for 1024 bit RSA signs.
17004 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17005 handling. Most clients have the effective key size in bits equal to
17006 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17007 A few however don't do this and instead use the size of the decrypted key
17008 to determine the RC2 key length and the AlgorithmIdentifier to determine
17009 the effective key length. In this case the effective key length can still
17010 be 40 bits but the key length can be 168 bits for example. This is fixed
17011 by manually forcing an RC2 key into the EVP_PKEY structure because the
17012 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17013 the key length and effective key length are equal.
17017 * Add a bunch of functions that should simplify the creation of
17018 X509_NAME structures. Now you should be able to do:
17019 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17020 and have it automatically work out the correct field type and fill in
17021 the structures. The more adventurous can try:
17022 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17023 and it will (hopefully) work out the correct multibyte encoding.
17027 * Change the 'req' utility to use the new field handling and multibyte
17028 copy routines. Before the DN field creation was handled in an ad hoc
17029 way in req, ca, and x509 which was rather broken and didn't support
17030 BMPStrings or UTF8Strings. Since some software doesn't implement
17031 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17032 using the dirstring_type option. See the new comment in the default
17033 openssl.cnf for more info.
17037 * Make crypto/rand/md_rand.c more robust:
17038 - Assure unique random numbers after fork().
17039 - Make sure that concurrent threads access the global counter and
17040 md serializably so that we never lose entropy in them
17041 or use exactly the same state in multiple threads.
17042 Access to the large state is not always serializable because
17043 the additional locking could be a performance killer, and
17044 md should be large enough anyway.
17048 * New file `apps/app_rand.c` with commonly needed functionality
17049 for handling the random seed file.
17051 Use the random seed file in some applications that previously did not:
17053 dsaparam -genkey (which also ignored its '-rand' option),
17056 x509 (when signing).
17057 Except on systems with /dev/urandom, it is crucial to have a random
17058 seed file at least for key creation, DSA signing, and for DH exchanges;
17059 for RSA signatures we could do without one.
17061 gendh and gendsa (unlike genrsa) used to read only the first byte
17062 of each file listed in the '-rand' option. The function as previously
17063 found in genrsa is now in app_rand.c and is used by all programs
17064 that support '-rand'.
17068 * In RAND_write_file, use mode 0600 for creating files;
17069 don't just chmod when it may be too late.
17073 * Report an error from X509_STORE_load_locations
17074 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17078 * New function ASN1_mbstring_copy() this copies a string in either
17079 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17080 into an ASN1_STRING type. A mask of permissible types is passed
17081 and it chooses the "minimal" type to use or an error if not type
17086 * Add function equivalents to the various macros in asn1.h. The old
17087 macros are retained with an `M_` prefix. Code inside the library can
17088 use the `M_` macros. External code (including the openssl utility)
17089 should *NOT* in order to be "shared library friendly".
17093 * Add various functions that can check a certificate's extensions
17094 to see if it usable for various purposes such as SSL client,
17095 server or S/MIME and CAs of these types. This is currently
17096 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17097 verification. Also added a -purpose flag to x509 utility to
17098 print out all the purposes.
17102 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17107 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17108 for, obtain and decode and extension and obtain its critical flag.
17109 This allows all the necessary extension code to be handled in a
17110 single function call.
17114 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17115 platforms. See crypto/rc4/rc4_enc.c for further details.
17119 * New -noout option to asn1parse. This causes no output to be produced
17120 its main use is when combined with -strparse and -out to extract data
17121 from a file (which may not be in ASN.1 format).
17125 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17126 when producing the local key id.
17128 *Richard Levitte <levitte@stacken.kth.se>*
17130 * New option -dhparam in s_server. This allows a DH parameter file to be
17131 stated explicitly. If it is not stated then it tries the first server
17132 certificate file. The previous behaviour hard coded the filename
17137 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17138 a public key to be input or output. For example:
17139 openssl rsa -in key.pem -pubout -out pubkey.pem
17140 Also added necessary DSA public key functions to handle this.
17144 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17145 in the message. This was handled by allowing
17146 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17148 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17150 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17151 to the end of the strings whereas this didn't. This would cause problems
17152 if strings read with d2i_ASN1_bytes() were later modified.
17154 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17156 * Fix for base64 decode bug. When a base64 bio reads only one line of
17157 data and it contains EOF it will end up returning an error. This is
17158 caused by input 46 bytes long. The cause is due to the way base64
17159 BIOs find the start of base64 encoded data. They do this by trying a
17160 trial decode on each line until they find one that works. When they
17161 do a flag is set and it starts again knowing it can pass all the
17162 data directly through the decoder. Unfortunately it doesn't reset
17163 the context it uses. This means that if EOF is reached an attempt
17164 is made to pass two EOFs through the context and this causes the
17165 resulting error. This can also cause other problems as well. As is
17166 usual with these problems it takes *ages* to find and the fix is
17167 trivial: move one line.
17169 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17171 * Ugly workaround to get s_client and s_server working under Windows. The
17172 old code wouldn't work because it needed to select() on sockets and the
17173 tty (for keypresses and to see if data could be written). Win32 only
17174 supports select() on sockets so we select() with a 1s timeout on the
17175 sockets and then see if any characters are waiting to be read, if none
17176 are present then we retry, we also assume we can always write data to
17177 the tty. This isn't nice because the code then blocks until we've
17178 received a complete line of data and it is effectively polling the
17179 keyboard at 1s intervals: however it's quite a bit better than not
17180 working at all :-) A dedicated Windows application might handle this
17181 with an event loop for example.
17185 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17186 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17187 will be called when RSA_sign() and RSA_verify() are used. This is useful
17188 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17189 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17190 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17191 This necessitated the support of an extra signature type NID_md5_sha1
17192 for SSL signatures and modifications to the SSL library to use it instead
17193 of calling RSA_public_decrypt() and RSA_private_encrypt().
17197 * Add new -verify -CAfile and -CApath options to the crl program, these
17198 will lookup a CRL issuers certificate and verify the signature in a
17199 similar way to the verify program. Tidy up the crl program so it
17200 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17201 less strict. It will now permit CRL extensions even if it is not
17202 a V2 CRL: this will allow it to tolerate some broken CRLs.
17206 * Initialize all non-automatic variables each time one of the openssl
17207 sub-programs is started (this is necessary as they may be started
17208 multiple times from the "OpenSSL>" prompt).
17210 *Lennart Bang, Bodo Moeller*
17212 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17213 removing all other RSA functionality (this is what NO_RSA does). This
17214 is so (for example) those in the US can disable those operations covered
17215 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17220 * Non-copying interface to BIO pairs.
17221 (still largely untested)
17225 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17226 ASCII string. This was handled independently in various places before.
17230 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17231 UTF8 strings a character at a time.
17235 * Use client_version from client hello to select the protocol
17236 (s23_srvr.c) and for RSA client key exchange verification
17237 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17241 * Add various utility functions to handle SPKACs, these were previously
17242 handled by poking round in the structure internals. Added new function
17243 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17244 print, verify and generate SPKACs. Based on an original idea from
17245 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17249 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17253 * Allow the config file extension section to be overwritten on the
17254 command line. Based on an original idea from Massimiliano Pala
17255 <madwolf@comune.modena.it>. The new option is called -extensions
17256 and can be applied to ca, req and x509. Also -reqexts to override
17257 the request extensions in req and -crlexts to override the crl extensions
17262 * Add new feature to the SPKAC handling in ca. Now you can include
17263 the same field multiple times by preceding it by "XXXX." for example:
17266 this is the same syntax as used in the req config file.
17270 * Allow certificate extensions to be added to certificate requests. These
17271 are specified in a 'req_extensions' option of the req section of the
17272 config file. They can be printed out with the -text option to req but
17273 are otherwise ignored at present.
17277 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17278 data read consists of only the final block it would not decrypted because
17279 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17280 A misplaced 'break' also meant the decrypted final block might not be
17281 copied until the next read.
17285 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17286 a few extra parameters to the DH structure: these will be useful if
17287 for example we want the value of 'q' or implement X9.42 DH.
17291 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17292 provides hooks that allow the default DSA functions or functions on a
17293 "per key" basis to be replaced. This allows hardware acceleration and
17294 hardware key storage to be handled without major modification to the
17295 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17296 associated functions.
17300 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17301 as "read only": it can't be written to and the buffer it points to will
17302 not be freed. Reading from a read only BIO is much more efficient than
17303 a normal memory BIO. This was added because there are several times when
17304 an area of memory needs to be read from a BIO. The previous method was
17305 to create a memory BIO and write the data to it, this results in two
17306 copies of the data and an O(n^2) reading algorithm. There is a new
17307 function BIO_new_mem_buf() which creates a read only memory BIO from
17308 an area of memory. Also modified the PKCS#7 routines to use read only
17313 * Bugfix: ssl23_get_client_hello did not work properly when called in
17314 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17315 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17316 but a retry condition occurred while trying to read the rest.
17320 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17321 NID_pkcs7_encrypted by default: this was wrong since this should almost
17322 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17323 the encrypted data type: this is a more sensible place to put it and it
17324 allows the PKCS#12 code to be tidied up that duplicated this
17329 * Changed obj_dat.pl script so it takes its input and output files on
17330 the command line. This should avoid shell escape redirection problems
17335 * Initial support for certificate extension requests, these are included
17336 in things like Xenroll certificate requests. Included functions to allow
17337 extensions to be obtained and added.
17341 * -crlf option to s_client and s_server for sending newlines as
17342 CRLF (as required by many protocols).
17346 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17348 * Install libRSAglue.a when OpenSSL is built with RSAref.
17350 *Ralf S. Engelschall*
17352 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17354 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17356 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17361 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17362 DH parameters/keys (q is lost during that conversion, but the resulting
17363 DH parameters contain its length).
17365 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17366 much faster than DH_generate_parameters (which creates parameters
17367 where `p = 2*q + 1`), and also the smaller q makes DH computations
17368 much more efficient (160-bit exponentiation instead of 1024-bit
17369 exponentiation); so this provides a convenient way to support DHE
17370 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17371 utter importance to use
17372 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17374 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17375 when such DH parameters are used, because otherwise small subgroup
17376 attacks may become possible!
17380 * Avoid memory leak in i2d_DHparams.
17384 * Allow the -k option to be used more than once in the enc program:
17385 this allows the same encrypted message to be read by multiple recipients.
17389 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17390 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17391 it will always use the numerical form of the OID, even if it has a short
17396 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17397 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17398 otherwise bn_mod_exp was called. In the case of hardware keys for example
17399 no private key components need be present and it might store extra data
17400 in the RSA structure, which cannot be accessed from bn_mod_exp.
17401 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17402 private key operations.
17406 * Added support for SPARC Linux.
17410 * pem_password_cb function type incompatibly changed from
17411 typedef int pem_password_cb(char *buf, int size, int rwflag);
17413 ....(char *buf, int size, int rwflag, void *userdata);
17414 so that applications can pass data to their callbacks:
17415 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17416 additional void * argument, which is just handed through whenever
17417 the password callback is called.
17419 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17421 New function SSL_CTX_set_default_passwd_cb_userdata.
17423 Compatibility note: As many C implementations push function arguments
17424 onto the stack in reverse order, the new library version is likely to
17425 interoperate with programs that have been compiled with the old
17426 pem_password_cb definition (PEM_whatever takes some data that
17427 happens to be on the stack as its last argument, and the callback
17428 just ignores this garbage); but there is no guarantee whatsoever that
17431 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17432 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17433 problems not only on Windows, but also on some Unix platforms.
17434 To avoid problematic command lines, these definitions are now in an
17435 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17436 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17440 * MIPS III/IV assembler module is reimplemented.
17444 * More DES library cleanups: remove references to srand/rand and
17445 delete an unused file.
17449 * Add support for the free Netwide assembler (NASM) under Win32,
17450 since not many people have MASM (ml) and it can be hard to obtain.
17451 This is currently experimental but it seems to work OK and pass all
17452 the tests. Check out INSTALL.W32 for info.
17456 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17457 without temporary keys kept an extra copy of the server key,
17458 and connections with temporary keys did not free everything in case
17463 * New function RSA_check_key and new openssl rsa option -check
17464 for verifying the consistency of RSA keys.
17466 *Ulf Moeller, Bodo Moeller*
17468 * Various changes to make Win32 compile work:
17469 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17470 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17471 comparison" warnings.
17472 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17476 * Add a debugging option to PKCS#5 v2 key generation function: when
17477 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17478 derived keys are printed to stderr.
17482 * Copy the flags in ASN1_STRING_dup().
17484 *Roman E. Pavlov <pre@mo.msk.ru>*
17486 * The x509 application mishandled signing requests containing DSA
17487 keys when the signing key was also DSA and the parameters didn't match.
17489 It was supposed to omit the parameters when they matched the signing key:
17490 the verifying software was then supposed to automatically use the CA's
17491 parameters if they were absent from the end user certificate.
17493 Omitting parameters is no longer recommended. The test was also
17494 the wrong way round! This was probably due to unusual behaviour in
17495 EVP_cmp_parameters() which returns 1 if the parameters match.
17496 This meant that parameters were omitted when they *didn't* match and
17497 the certificate was useless. Certificates signed with 'ca' didn't have
17500 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17502 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17503 The interface is as follows:
17504 Applications can use
17505 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17506 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17507 "off" is now the default.
17508 The library internally uses
17509 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17510 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17511 to disable memory-checking temporarily.
17513 Some inconsistent states that previously were possible (and were
17514 even the default) are now avoided.
17516 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17517 with each memory chunk allocated; this is occasionally more helpful
17518 than just having a counter.
17520 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17522 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17527 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17528 which largely parallels "options", but is for changing API behaviour,
17529 whereas "options" are about protocol behaviour.
17530 Initial "mode" flags are:
17532 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17533 a single record has been written.
17534 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17535 retries use the same buffer location.
17536 (But all of the contents must be
17541 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17544 * Fix problems with no-hmac etc.
17546 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17548 * New functions RSA_get_default_method(), RSA_set_method() and
17549 RSA_get_method(). These allows replacement of RSA_METHODs without having
17550 to mess around with the internals of an RSA structure.
17554 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17555 Also really enable memory leak checks in openssl.c and in some
17558 *Chad C. Mulligan, Bodo Moeller*
17560 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17561 up the length of negative integers. This has now been simplified to just
17562 store the length when it is first determined and use it later, rather
17563 than trying to keep track of where data is copied and updating it to
17565 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17567 * Add a new function PKCS7_signatureVerify. This allows the verification
17568 of a PKCS#7 signature but with the signing certificate passed to the
17569 function itself. This contrasts with PKCS7_dataVerify which assumes the
17570 certificate is present in the PKCS#7 structure. This isn't always the
17571 case: certificates can be omitted from a PKCS#7 structure and be
17572 distributed by "out of band" means (such as a certificate database).
17576 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17577 function prototypes in pem.h, also change util/mkdef.pl to add the
17578 necessary function names.
17582 * mk1mf.pl (used by Windows builds) did not properly read the
17583 options set by Configure in the top level Makefile, and Configure
17584 was not even able to write more than one option correctly.
17585 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17589 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17590 file to be loaded from a BIO or FILE pointer. The BIO version will
17591 for example allow memory BIOs to contain config info.
17595 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17596 Whoever hopes to achieve shared-library compatibility across versions
17597 must use this, not the compile-time macro.
17598 (Exercise 0.9.4: Which is the minimum library version required by
17600 Note: All this applies only to multi-threaded programs, others don't
17605 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17606 through a BIO pair triggered the default case, i.e.
17607 SSLerr(...,SSL_R_UNKNOWN_STATE).
17611 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17612 can use the SSL library even if none of the specific BIOs is
17617 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17618 for the encoded length.
17620 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17622 * Add initial documentation of the X509V3 functions.
17626 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17627 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17628 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17629 secure PKCS#8 private key format with a high iteration count.
17633 * Fix determination of Perl interpreter: A perl or perl5
17634 *directory* in $PATH was also accepted as the interpreter.
17636 *Ralf S. Engelschall*
17638 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17639 wrong with it but it was very old and did things like calling
17640 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17641 unusual formatting.
17645 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17646 to use the new extension code.
17650 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17651 with macros. This should make it easier to change their form, add extra
17652 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17657 * Add to configuration table a new entry that can specify an alternative
17658 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17659 according to Mark Crispin <MRC@Panda.COM>.
17663 * DES CBC did not update the IV. Weird.
17667 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17668 Changing the behaviour of the former might break existing programs --
17669 where IV updating is needed, des_ncbc_encrypt can be used.
17672 * When bntest is run from "make test" it drives bc to check its
17673 calculations, as well as internally checking them. If an internal check
17674 fails, it needs to cause bc to give a non-zero result or make test carries
17675 on without noticing the failure. Fixed.
17679 * DES library cleanups.
17683 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17684 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17685 ciphers. NOTE: although the key derivation function has been verified
17686 against some published test vectors it has not been extensively tested
17687 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17692 * Instead of "mkdir -p", which is not fully portable, use new
17693 Perl script "util/mkdir-p.pl".
17697 * Rewrite the way password based encryption (PBE) is handled. It used to
17698 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17699 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17700 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17701 the 'parameter' field of the AlgorithmIdentifier is passed to the
17702 underlying key generation function so it must do its own ASN1 parsing.
17703 This has also changed the EVP_PBE_CipherInit() function which now has a
17704 'parameter' argument instead of literal salt and iteration count values
17705 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17709 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17710 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17711 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17712 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17713 value was just used as a "magic string" and not used directly its
17714 value doesn't matter.
17718 * Introduce some semblance of const correctness to BN. Shame C doesn't
17723 * "linux-sparc64" configuration (ultrapenguin).
17725 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17726 "linux-sparc" configuration.
17728 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17730 * config now generates no-xxx options for missing ciphers.
17734 * Support the EBCDIC character set (work in progress).
17735 File ebcdic.c not yet included because it has a different license.
17737 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17739 * Support BS2000/OSD-POSIX.
17741 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17743 * Make callbacks for key generation use `void *` instead of `char *`.
17747 * Make S/MIME samples compile (not yet tested).
17751 * Additional typesafe stacks.
17755 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17759 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17761 * New configuration variant "sco5-gcc".
17763 * Updated some demos.
17765 *Sean O Riordain, Wade Scholine*
17767 * Add missing BIO_free at exit of pkcs12 application.
17771 * Fix memory leak in conf.c.
17775 * Updates for Win32 to assembler version of MD5.
17779 * Set #! path to perl in `apps/der_chop` to where we found it
17780 instead of using a fixed path.
17784 * SHA library changes for irix64-mips4-cc.
17788 * Improvements for VMS support.
17792 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17794 * Bignum library bug fix. IRIX 6 passes "make test" now!
17795 This also avoids the problems with SC4.2 and unpatched SC5.
17797 *Andy Polyakov <appro@fy.chalmers.se>*
17799 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17800 These are required because of the typesafe stack would otherwise break
17801 existing code. If old code used a structure member which used to be STACK
17802 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17803 sk_num or sk_value it would produce an error because the num, data members
17804 are not present in STACK_OF. Now it just produces a warning. sk_set
17805 replaces the old method of assigning a value to sk_value
17806 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17807 that does this will no longer work (and should use sk_set instead) but
17808 this could be regarded as a "questionable" behaviour anyway.
17812 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17813 correctly handle encrypted S/MIME data.
17817 * Change type of various DES function arguments from des_cblock
17818 (which means, in function argument declarations, pointer to char)
17819 to des_cblock * (meaning pointer to array with 8 char elements),
17820 which allows the compiler to do more typechecking; it was like
17821 that back in SSLeay, but with lots of ugly casts.
17823 Introduce new type const_des_cblock.
17827 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17828 problems: find RecipientInfo structure that matches recipient certificate
17829 and initialise the ASN1 structures properly based on passed cipher.
17833 * Belatedly make the BN tests actually check the results.
17837 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17838 to and from BNs: it was completely broken. New compilation option
17839 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17840 key elements as negative integers.
17844 * Reorganize and speed up MD5.
17846 *Andy Polyakov <appro@fy.chalmers.se>*
17850 *Richard Levitte <richard@levitte.org>*
17852 * New option -out to asn1parse to allow the parsed structure to be
17853 output to a file. This is most useful when combined with the -strparse
17854 option to examine the output of things like OCTET STRINGS.
17858 * Make SSL library a little more fool-proof by not requiring any longer
17859 that `SSL_set_{accept,connect}_state` be called before
17860 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17861 in many applications because usually everything *appeared* to work as
17862 intended anyway -- now it really works as intended).
17866 * Move openssl.cnf out of lib/.
17870 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17871 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17872 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17874 *Ralf S. Engelschall*
17876 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17877 handle PKCS#7 enveloped data properly.
17879 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17881 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17882 copying pointers. The cert_st handling is changed by this in
17883 various ways (and thus what used to be known as ctx->default_cert
17884 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17885 any longer when s->cert does not give us what we need).
17886 ssl_cert_instantiate becomes obsolete by this change.
17887 As soon as we've got the new code right (possibly it already is?),
17888 we have solved a couple of bugs of the earlier code where s->cert
17889 was used as if it could not have been shared with other SSL structures.
17891 Note that using the SSL API in certain dirty ways now will result
17892 in different behaviour than observed with earlier library versions:
17893 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17894 does not influence s as it used to.
17896 In order to clean up things more thoroughly, inside SSL_SESSION
17897 we don't use CERT any longer, but a new structure SESS_CERT
17898 that holds per-session data (if available); currently, this is
17899 the peer's certificate chain and, for clients, the server's certificate
17900 and temporary key. CERT holds only those values that can have
17901 meaningful defaults in an SSL_CTX.
17905 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17906 from the internal representation. Various PKCS#7 fixes: remove some
17907 evil casts and set the enc_dig_alg field properly based on the signing
17912 * Allow PKCS#12 password to be set from the command line or the
17913 environment. Let 'ca' get its config file name from the environment
17914 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17919 * Allow certificate policies extension to use an IA5STRING for the
17920 organization field. This is contrary to the PKIX definition but
17921 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17926 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17927 without disallowing inline assembler and the like for non-pedantic builds.
17931 * Support Borland C++ builder.
17933 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17939 * SHA-1 cleanups and performance enhancements.
17941 *Andy Polyakov <appro@fy.chalmers.se>*
17943 * Sparc v8plus assembler for the bignum library.
17945 *Andy Polyakov <appro@fy.chalmers.se>*
17947 * Accept any -xxx and +xxx compiler options in Configure.
17951 * Update HPUX configuration.
17955 * Add missing `sk_<type>_unshift()` function to safestack.h
17957 *Ralf S. Engelschall*
17959 * New function SSL_CTX_use_certificate_chain_file that sets the
17960 "extra_cert"s in addition to the certificate. (This makes sense
17961 only for "PEM" format files, as chains as a whole are not
17966 * Support verify_depth from the SSL API.
17967 x509_vfy.c had what can be considered an off-by-one-error:
17968 Its depth (which was not part of the external interface)
17969 was actually counting the number of certificates in a chain;
17970 now it really counts the depth.
17974 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17975 instead of X509err, which often resulted in confusing error
17976 messages since the error codes are not globally unique
17977 (e.g. an alleged error in ssl3_accept when a certificate
17978 didn't match the private key).
17980 * New function SSL_CTX_set_session_id_context that allows to set a default
17981 value (so that you don't need SSL_set_session_id_context for each
17982 connection using the SSL_CTX).
17986 * OAEP decoding bug fix.
17990 * Support INSTALL_PREFIX for package builders, as proposed by
17995 * New Configure options "threads" and "no-threads". For systems
17996 where the proper compiler options are known (currently Solaris
17997 and Linux), "threads" is the default.
18001 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18005 * Install various scripts to $(OPENSSLDIR)/misc, not to
18006 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18007 such as /usr/local/bin.
18011 * "make linux-shared" to build shared libraries.
18013 *Niels Poppe <niels@netbox.org>*
18015 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18019 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18020 extension adding in x509 utility.
18024 * Remove NOPROTO sections and error code comments.
18028 * Partial rewrite of the DEF file generator to now parse the ANSI
18033 * New Configure options --prefix=DIR and --openssldir=DIR.
18037 * Complete rewrite of the error code script(s). It is all now handled
18038 by one script at the top level which handles error code gathering,
18039 header rewriting and C source file generation. It should be much better
18040 than the old method: it now uses a modified version of Ulf's parser to
18041 read the ANSI prototypes in all header files (thus the old K&R definitions
18042 aren't needed for error creation any more) and do a better job of
18043 translating function codes into names. The old 'ASN1 error code embedded
18044 in a comment' is no longer necessary and it doesn't use .err files which
18045 have now been deleted. Also the error code call doesn't have to appear all
18046 on one line (which resulted in some large lines...).
18050 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18054 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18055 0 (which usually indicates a closed connection), but continue reading.
18059 * Fix some race conditions.
18063 * Add support for CRL distribution points extension. Add Certificate
18064 Policies and CRL distribution points documentation.
18068 * Move the autogenerated header file parts to crypto/opensslconf.h.
18072 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18073 8 of keying material. Merlin has also confirmed interop with this fix
18074 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18076 *Merlin Hughes <merlin@baltimore.ie>*
18078 * Fix lots of warnings.
18080 *Richard Levitte <levitte@stacken.kth.se>*
18082 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18083 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18085 *Richard Levitte <levitte@stacken.kth.se>*
18087 * Fix problems with sizeof(long) == 8.
18089 *Andy Polyakov <appro@fy.chalmers.se>*
18091 * Change functions to ANSI C.
18095 * Fix typos in error codes.
18097 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18099 * Remove defunct assembler files from Configure.
18103 * SPARC v8 assembler BIGNUM implementation.
18105 *Andy Polyakov <appro@fy.chalmers.se>*
18107 * Support for Certificate Policies extension: both print and set.
18108 Various additions to support the r2i method this uses.
18112 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18113 return a const string when you are expecting an allocated buffer.
18117 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18118 types DirectoryString and DisplayText.
18122 * Add code to allow r2i extensions to access the configuration database,
18123 add an LHASH database driver and add several ctx helper functions.
18127 * Fix an evil bug in bn_expand2() which caused various BN functions to
18128 fail when they extended the size of a BIGNUM.
18132 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18133 support typesafe stack.
18137 * Fix typo in SSL_[gs]et_options().
18139 *Nils Frostberg <nils@medcom.se>*
18141 * Delete various functions and files that belonged to the (now obsolete)
18142 old X509V3 handling code.
18146 * New Configure option "rsaref".
18150 * Don't auto-generate pem.h.
18154 * Introduce type-safe ASN.1 SETs.
18158 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18160 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18162 * Introduce type-safe STACKs. This will almost certainly break lots of code
18163 that links with OpenSSL (well at least cause lots of warnings), but fear
18164 not: the conversion is trivial, and it eliminates loads of evil casts. A
18165 few STACKed things have been converted already. Feel free to convert more.
18166 In the fullness of time, I'll do away with the STACK type altogether.
18170 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18171 specified in `<certfile>` by updating the entry in the index.txt file.
18172 This way one no longer has to edit the index.txt file manually for
18173 revoking a certificate. The -revoke option does the gory details now.
18175 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18177 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18178 `-text` option at all and this way the `-noout -text` combination was
18179 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18181 *Ralf S. Engelschall*
18183 * Make sure a corresponding plain text error message exists for the
18184 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18185 verify callback function determined that a certificate was revoked.
18187 *Ralf S. Engelschall*
18189 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18190 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18191 all available ciphers including rc5, which was forgotten until now.
18192 In order to let the testing shell script know which algorithms
18193 are available, a new (up to now undocumented) command
18194 `openssl list-cipher-commands` is used.
18198 * Bugfix: s_client occasionally would sleep in select() when
18199 it should have checked SSL_pending() first.
18203 * New functions DSA_do_sign and DSA_do_verify to provide access to
18204 the raw DSA values prior to ASN.1 encoding.
18208 * Tweaks to Configure
18210 *Niels Poppe <niels@netbox.org>*
18212 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18217 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18221 * New config option to avoid instructions that are illegal on the 80386.
18222 The default code is faster, but requires at least a 486.
18226 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18227 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18228 same as SSL2_VERSION anyway.
18232 * New "-showcerts" option for s_client.
18236 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18237 application. Various cleanups and fixes.
18241 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18242 modify error routines to work internally. Add error codes and PBE init
18243 to library startup routines.
18247 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18248 packing functions to asn1 and evp. Changed function names and error
18249 codes along the way.
18253 * PKCS12 integration: and so it begins... First of several patches to
18254 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18255 objects to objects.h
18259 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18260 and display support for Thawte strong extranet extension.
18264 * Add LinuxPPC support.
18266 *Jeff Dubrule <igor@pobox.org>*
18268 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18269 bn_div_words in alpha.s.
18271 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18273 * Make sure the RSA OAEP test is skipped under -DRSAref because
18274 OAEP isn't supported when OpenSSL is built with RSAref.
18276 *Ulf Moeller <ulf@fitug.de>*
18278 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18279 so they no longer are missing under -DNOPROTO.
18281 *Soren S. Jorvang <soren@t.dk>*
18283 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18285 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18286 doesn't work when the session is reused. Coming soon!
18290 * Fix a security hole, that allows sessions to be reused in the wrong
18291 context thus bypassing client cert protection! All software that uses
18292 client certs and session caches in multiple contexts NEEDS PATCHING to
18293 allow session reuse! A fuller solution is in the works.
18295 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18297 * Some more source tree cleanups (removed obsolete files
18298 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18299 permission on "config" script to be executable) and a fix for the INSTALL
18302 *Ulf Moeller <ulf@fitug.de>*
18304 * Remove some legacy and erroneous uses of malloc, free instead of
18307 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18309 * Make rsa_oaep_test return non-zero on error.
18311 *Ulf Moeller <ulf@fitug.de>*
18313 * Add support for native Solaris shared libraries. Configure
18314 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18315 if someone would make that last step automatic.
18317 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18319 * ctx_size was not built with the right compiler during "make links". Fixed.
18323 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18324 except NULL ciphers". This means the default cipher list will no longer
18325 enable NULL ciphers. They need to be specifically enabled e.g. with
18326 the string "DEFAULT:eNULL".
18330 * Fix to RSA private encryption routines: if p < q then it would
18331 occasionally produce an invalid result. This will only happen with
18332 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18336 * Be less restrictive and allow also `perl util/perlpath.pl
18337 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18338 because this way one can also use an interpreter named `perl5` (which is
18339 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18340 installed as `perl`).
18342 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18344 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18346 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18348 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18349 advapi32.lib to Win32 build and change the pem test comparison
18350 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18351 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18352 and crypto/des/ede_cbcm_enc.c.
18356 * DES quad checksum was broken on big-endian architectures. Fixed.
18360 * Comment out two functions in bio.h that aren't implemented. Fix up the
18361 Win32 test batch file so it (might) work again. The Win32 test batch file
18362 is horrible: I feel ill....
18366 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18367 in e_os.h. Audit of header files to check ANSI and non ANSI
18368 sections: 10 functions were absent from non ANSI section and not exported
18369 from Windows DLLs. Fixed up libeay.num for new functions.
18373 * Make `openssl version` output lines consistent.
18375 *Ralf S. Engelschall*
18377 * Fix Win32 symbol export lists for BIO functions: Added
18378 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18379 to ms/libeay{16,32}.def.
18381 *Ralf S. Engelschall*
18383 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18384 fine under Unix and passes some trivial tests I've now added. But the
18385 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18386 added to make sure no one expects that this stuff really works in the
18387 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18388 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18391 *Ralf S. Engelschall*
18393 * Fix the generation of two part addresses in perl.
18395 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18397 * Add config entry for Linux on MIPS.
18399 *John Tobey <jtobey@channel1.com>*
18401 * Make links whenever Configure is run, unless we are on Windoze.
18405 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18406 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18411 * Add a useful kludge to allow package maintainers to specify compiler and
18412 other platforms details on the command line without having to patch the
18413 Configure script every time: One now can use
18414 `perl Configure <id>:<details>`,
18415 i.e. platform ids are allowed to have details appended
18416 to them (separated by colons). This is treated as there would be a static
18417 pre-configured entry in Configure's %table under key `<id>` with value
18418 `<details>` and `perl Configure <id>` is called. So, when you want to
18419 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18420 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18421 now, which overrides the FreeBSD-elf entry on-the-fly.
18423 *Ralf S. Engelschall*
18425 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18429 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18430 on the `perl Configure ...` command line. This way one can compile
18431 OpenSSL libraries with Position Independent Code (PIC) which is needed
18432 for linking it into DSOs.
18434 *Ralf S. Engelschall*
18436 * Remarkably, export ciphers were totally broken and no-one had noticed!
18441 * Cleaned up the LICENSE document: The official contact for any license
18442 questions now is the OpenSSL core team under openssl-core@openssl.org.
18443 And add a paragraph about the dual-license situation to make sure people
18444 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18445 to the OpenSSL toolkit.
18447 *Ralf S. Engelschall*
18449 * General source tree makefile cleanups: Made `making xxx in yyy...`
18450 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18451 Additionally cleaned up the `make links` target: Remove unnecessary
18452 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18453 to speed processing and no longer clutter the display with confusing
18454 stuff. Instead only the actually done links are displayed.
18456 *Ralf S. Engelschall*
18458 * Permit null encryption ciphersuites, used for authentication only. It used
18459 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18460 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18465 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18466 signed attributes when verifying signatures (this would break them),
18467 the detached data encoding was wrong and public keys obtained using
18468 X509_get_pubkey() weren't freed.
18472 * Add text documentation for the BUFFER functions. Also added a work around
18473 to a Win95 console bug. This was triggered by the password read stuff: the
18474 last character typed gets carried over to the next fread(). If you were
18475 generating a new cert request using 'req' for example then the last
18476 character of the passphrase would be CR which would then enter the first
18481 * Added the new 'Includes OpenSSL Cryptography Software' button as
18482 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18483 button and can be used by applications based on OpenSSL to show the
18484 relationship to the OpenSSL project.
18486 *Ralf S. Engelschall*
18488 * Remove confusing variables in function signatures in files
18489 ssl/ssl_lib.c and ssl/ssl.h.
18491 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18493 * Don't install bss_file.c under PREFIX/include/
18495 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18497 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18498 functions that return function pointers and has support for NT specific
18499 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18500 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18501 unsigned to signed types: this was killing the Win32 compile.
18505 * Add new certificate file to stack functions,
18506 SSL_add_dir_cert_subjects_to_stack() and
18507 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18508 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18509 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18510 This means that Apache-SSL and similar packages don't have to mess around
18511 to add as many CAs as they want to the preferred list.
18515 * Experiment with doxygen documentation. Currently only partially applied to
18517 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18518 openssl.doxy as the configuration file.
18522 * Get rid of remaining C++-style comments which strict C compilers hate.
18524 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18526 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18527 compiled in by default: it has problems with large keys.
18531 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18532 DH private keys and/or callback functions which directly correspond to
18533 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18534 is needed for applications which have to configure certificates on a
18535 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18537 For the RSA certificate situation is makes no difference, but
18538 for the DSA certificate situation this fixes the "no shared cipher"
18539 problem where the OpenSSL cipher selection procedure failed because the
18540 temporary keys were not overtaken from the context and the API provided
18541 no way to reconfigure them.
18542 The new functions now let applications reconfigure the stuff and they
18543 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18544 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18545 non-public-API function ssl_cert_instantiate() is used as a helper
18546 function and also to reduce code redundancy inside ssl_rsa.c.
18548 *Ralf S. Engelschall*
18550 * Move s_server -dcert and -dkey options out of the undocumented feature
18551 area because they are useful for the DSA situation and should be
18552 recognized by the users.
18554 *Ralf S. Engelschall*
18556 * Fix the cipher decision scheme for export ciphers: the export bits are
18557 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18558 SSL_EXP_MASK. So, the original variable has to be used instead of the
18559 already masked variable.
18561 *Richard Levitte <levitte@stacken.kth.se>*
18563 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18565 *Richard Levitte <levitte@stacken.kth.se>*
18567 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18568 from `int` to `unsigned int` because it is a length and initialized by
18569 EVP_DigestFinal() which expects an `unsigned int *`.
18571 *Richard Levitte <levitte@stacken.kth.se>*
18573 * Don't hard-code path to Perl interpreter on shebang line of Configure
18574 script. Instead use the usual Shell->Perl transition trick.
18576 *Ralf S. Engelschall*
18578 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18579 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18580 -noout -modulus` as it's already the case for `openssl rsa -noout
18581 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18582 currently the public key is printed (a decision which was already done by
18583 `openssl dsa -modulus` in the past) which serves a similar purpose.
18584 Additionally the NO_RSA no longer completely removes the whole -modulus
18585 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18588 *Ralf S. Engelschall*
18590 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18591 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18593 *Arne Ansper <arne@ats.cyber.ee>*
18595 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18596 to be added. Now both 'req' and 'ca' can use new objects defined in the
18601 * Add cool BIO that does syslog (or event log on NT).
18603 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18605 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18606 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18607 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18608 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18612 * Add preliminary config info for new extension code.
18616 * Make RSA_NO_PADDING really use no padding.
18618 *Ulf Moeller <ulf@fitug.de>*
18620 * Generate errors when private/public key check is done.
18624 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18625 for some CRL extensions and new objects added.
18629 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18630 key usage extension and fuller support for authority key id.
18634 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18635 padding method for RSA, which is recommended for new applications in PKCS
18636 #1 v2.0 (RFC 2437, October 1998).
18637 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18638 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18639 against Bleichbacher's attack on RSA.
18640 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18643 * Updates to the new SSL compression code
18645 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18647 * Fix so that the version number in the master secret, when passed
18648 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18649 (because the server will not accept higher), that the version number
18650 is 0x03,0x01, not 0x03,0x00
18652 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18654 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18655 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18656 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18660 * Support for RAW extensions where an arbitrary extension can be
18661 created by including its DER encoding. See `apps/openssl.cnf` for
18666 * Make sure latest Perl versions don't interpret some generated C array
18667 code as Perl array code in the crypto/err/err_genc.pl script.
18669 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18671 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18672 not many people have the assembler. Various Win32 compilation fixes and
18673 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18674 build instructions.
18678 * Modify configure script 'Configure' to automatically create crypto/date.h
18679 file under Win32 and also build pem.h from pem.org. New script
18680 util/mkfiles.pl to create the MINFO file on environments that can't do a
18681 'make files': perl util/mkfiles.pl >MINFO should work.
18685 * Major rework of DES function declarations, in the pursuit of correctness
18686 and purity. As a result, many evil casts evaporated, and some weirdness,
18687 too. You may find this causes warnings in your code. Zapping your evil
18688 casts will probably fix them. Mostly.
18692 * Fix for a typo in asn1.h. Bug fix to object creation script
18693 obj_dat.pl. It considered a zero in an object definition to mean
18694 "end of object": none of the objects in objects.h have any zeros
18695 so it wasn't spotted.
18697 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18699 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18700 Masking (CBCM). In the absence of test vectors, the best I have been able
18701 to do is check that the decrypt undoes the encrypt, so far. Send me test
18702 vectors if you have them.
18706 * Correct calculation of key length for export ciphers (too much space was
18707 allocated for null ciphers). This has not been tested!
18711 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18712 message is now correct (it understands "crypto" and "ssl" on its
18713 command line). There is also now an "update" option. This will update
18714 the util/ssleay.num and util/libeay.num files with any new functions.
18716 perl util/mkdef.pl crypto ssl update
18717 it will update them.
18721 * Overhauled the Perl interface:
18722 - ported BN stuff to OpenSSL's different BN library
18723 - made the perl/ source tree CVS-aware
18724 - renamed the package from SSLeay to OpenSSL (the files still contain
18725 their history because I've copied them in the repository)
18726 - removed obsolete files (the test scripts will be replaced
18727 by better Test::Harness variants in the future)
18729 *Ralf S. Engelschall*
18731 * First cut for a very conservative source tree cleanup:
18732 1. merge various obsolete readme texts into doc/ssleay.txt
18733 where we collect the old documents and readme texts.
18734 2. remove the first part of files where I'm already sure that we no
18735 longer need them because of three reasons: either they are just temporary
18736 files which were left by Eric or they are preserved original files where
18737 I've verified that the diff is also available in the CVS via "cvs diff
18738 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18739 the crypto/md/ stuff).
18741 *Ralf S. Engelschall*
18743 * More extension code. Incomplete support for subject and issuer alt
18744 name, issuer and authority key id. Change the i2v function parameters
18745 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18746 what that's for :-) Fix to ASN1 macro which messed up
18747 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18751 * Preliminary support for ENUMERATED type. This is largely copied from the
18756 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18758 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18760 * Make sure `make rehash` target really finds the `openssl` program.
18762 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18764 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18765 like to hear about it if this slows down other processors.
18769 * Add CygWin32 platform information to Configure script.
18771 *Alan Batie <batie@aahz.jf.intel.com>*
18773 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18775 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18777 * New program nseq to manipulate netscape certificate sequences
18781 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18786 * Fixes to BN code. Previously the default was to define BN_RECURSION
18787 but the BN code had some problems that would cause failures when
18788 doing certificate verification and some other functions.
18790 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18792 * Add ASN1 and PEM code to support netscape certificate sequences.
18796 * Add ASN1 and PEM code to support netscape certificate sequences.
18800 * Add several PKIX and private extended key usage OIDs.
18804 * Modify the 'ca' program to handle the new extension code. Modify
18805 openssl.cnf for new extension format, add comments.
18809 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18810 and add a sample to openssl.cnf so req -x509 now adds appropriate
18815 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18816 error code, add initial support to X509_print() and x509 application.
18820 * Takes a deep breath and start adding X509 V3 extension support code. Add
18821 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18822 stuff is currently isolated and isn't even compiled yet.
18826 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18827 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18828 Removed the versions check from X509 routines when loading extensions:
18829 this allows certain broken certificates that don't set the version
18830 properly to be processed.
18834 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18835 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18836 can still be regenerated with "make depend".
18840 * Spelling mistake in C version of CAST-128.
18842 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18844 * Changes to the error generation code. The perl script err-code.pl
18845 now reads in the old error codes and retains the old numbers, only
18846 adding new ones if necessary. It also only changes the .err files if new
18847 codes are added. The makefiles have been modified to only insert errors
18848 when needed (to avoid needlessly modifying header files). This is done
18849 by only inserting errors if the .err file is newer than the auto generated
18850 C file. To rebuild all the error codes from scratch (the old behaviour)
18851 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18852 or delete all the .err files.
18856 * CAST-128 was incorrectly implemented for short keys. The C version has
18857 been fixed, but is untested. The assembler versions are also fixed, but
18858 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18859 to regenerate it if needed.
18860 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18861 Hagino <itojun@kame.net>*
18863 * File was opened incorrectly in randfile.c.
18865 *Ulf Möller <ulf@fitug.de>*
18867 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18868 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18869 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18870 al: it's just almost always a UTCTime. Note this patch adds new error
18871 codes so do a "make errors" if there are problems.
18875 * Correct Linux 1 recognition in config.
18877 *Ulf Möller <ulf@fitug.de>*
18879 * Remove pointless MD5 hash when using DSA keys in ca.
18881 *Anonymous <nobody@replay.com>*
18883 * Generate an error if given an empty string as a cert directory. Also
18884 generate an error if handed NULL (previously returned 0 to indicate an
18885 error, but didn't set one).
18887 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18889 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18893 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18894 parameters. This was causing a warning which killed off the Win32 compile.
18898 * Remove C++ style comments from crypto/bn/bn_local.h.
18900 *Neil Costigan <neil.costigan@celocom.com>*
18902 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18903 based on a text string, looking up short and long names and finally
18904 "dot" format. The "dot" format stuff didn't work. Added new function
18905 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18906 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18907 OID is not part of the table.
18911 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18912 X509_LOOKUP_by_alias().
18916 * Sort openssl functions by name.
18920 * Get the `gendsa` command working and add it to the `list` command. Remove
18921 encryption from sample DSA keys (in case anyone is interested the password
18926 * Make *all* `*_free` functions accept a NULL pointer.
18928 *Frans Heymans <fheymans@isaserver.be>*
18930 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18933 *Anonymous <nobody@replay.com>*
18935 * s_server should send the CAfile as acceptable CAs, not its own cert.
18937 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18939 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18941 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18943 * Temp key "for export" tests were wrong in s3_srvr.c.
18945 *Anonymous <nobody@replay.com>*
18947 * Add prototype for temp key callback functions
18948 SSL_CTX_set_tmp_{rsa,dh}_callback().
18952 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18953 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18957 * X509_name_add_entry() freed the wrong thing after an error.
18959 *Arne Ansper <arne@ats.cyber.ee>*
18961 * rsa_eay.c would attempt to free a NULL context.
18963 *Arne Ansper <arne@ats.cyber.ee>*
18965 * BIO_s_socket() had a broken should_retry() on Windoze.
18967 *Arne Ansper <arne@ats.cyber.ee>*
18969 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18971 *Arne Ansper <arne@ats.cyber.ee>*
18973 * Make sure the already existing X509_STORE->depth variable is initialized
18974 in X509_STORE_new(), but document the fact that this variable is still
18975 unused in the certificate verification process.
18977 *Ralf S. Engelschall*
18979 * Fix the various library and `apps/` files to free up pkeys obtained from
18980 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18984 * Fix reference counting in X509_PUBKEY_get(). This makes
18985 demos/maurice/example2.c work, amongst others, probably.
18987 *Steve Henson and Ben Laurie*
18989 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18990 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18991 are no longer created. This way we have a single and consistent command
18992 line interface `openssl <command>`, similar to `cvs <command>`.
18994 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18996 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18997 BIT STRING wrapper always have zero unused bits.
19001 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19005 * Make the top-level INSTALL documentation easier to understand.
19009 * Makefiles updated to exit if an error occurs in a sub-directory
19010 make (including if user presses ^C) [Paul Sutton]
19012 * Make Montgomery context stuff explicit in RSA data structure.
19016 * Fix build order of pem and err to allow for generated pem.h.
19020 * Fix renumbering bug in X509_NAME_delete_entry().
19024 * Enhanced the err-ins.pl script so it makes the error library number
19025 global and can add a library name. This is needed for external ASN1 and
19026 other error libraries.
19030 * Fixed sk_insert which never worked properly.
19034 * Fix ASN1 macros so they can handle indefinite length constructed
19035 EXPLICIT tags. Some non standard certificates use these: they can now
19040 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19041 into a single doc/ssleay.txt bundle. This way the information is still
19042 preserved but no longer messes up this directory. Now it's new room for
19043 the new set of documentation files.
19045 *Ralf S. Engelschall*
19047 * SETs were incorrectly DER encoded. This was a major pain, because they
19048 shared code with SEQUENCEs, which aren't coded the same. This means that
19049 almost everything to do with SETs or SEQUENCEs has either changed name or
19050 number of arguments.
19052 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19054 * Fix test data to work with the above.
19058 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19059 was already fixed by Eric for 0.9.1 it seems.
19061 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19063 * Autodetect FreeBSD3.
19067 * Fix various bugs in Configure. This affects the following platforms:
19071 unixware-2.0-pentium
19076 * Eliminate generated files from CVS. Reorder tests to regenerate files
19077 before they are needed.
19081 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19085 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19087 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19088 changed SSLeay to OpenSSL in version strings.
19090 *Ralf S. Engelschall*
19092 * Some fixups to the top-level documents.
19096 * Fixed the nasty bug where rsaref.h was not found under compile-time
19097 because the symlink to include/ was missing.
19099 *Ralf S. Engelschall*
19101 * Incorporated the popular no-RSA/DSA-only patches
19102 which allow to compile a RSA-free SSLeay.
19104 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19106 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19107 when "ssleay" is still not found.
19109 *Ralf S. Engelschall*
19111 * Added more platforms to Configure: Cray T3E, HPUX 11,
19113 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19115 * Updated the README file.
19117 *Ralf S. Engelschall*
19119 * Added various .cvsignore files in the CVS repository subdirs
19120 to make a "cvs update" really silent.
19122 *Ralf S. Engelschall*
19124 * Recompiled the error-definition header files and added
19125 missing symbols to the Win32 linker tables.
19127 *Ralf S. Engelschall*
19129 * Cleaned up the top-level documents;
19130 o new files: CHANGES and LICENSE
19131 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19132 o merged COPYRIGHT into LICENSE
19133 o removed obsolete TODO file
19134 o renamed MICROSOFT to INSTALL.W32
19136 *Ralf S. Engelschall*
19138 * Removed dummy files from the 0.9.1b source tree:
19139 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19140 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19141 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19142 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19143 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19145 *Ralf S. Engelschall*
19147 * Added various platform portability fixes.
19151 * The Genesis of the OpenSSL rpject:
19152 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19153 Young and Tim J. Hudson created while they were working for C2Net until
19156 *The OpenSSL Project*
19158 ### Changes between 0.9.0b and 0.9.1b [not released]
19160 * Updated a few CA certificates under certs/
19164 * Changed some BIGNUM api stuff.
19168 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19169 DGUX x86, Linux Alpha, etc.
19173 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19174 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19179 * Add -strparse option to asn1pars program which parses nested
19182 *Dr Stephen Henson <shenson@bigfoot.com>*
19184 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19188 * DSA fix for "ca" program.
19192 * Added "-genkey" option to "dsaparam" program.
19196 * Added RIPE MD160 (rmd160) message digest.
19200 * Added -a (all) option to "ssleay version" command.
19204 * Added PLATFORM define which is the id given to Configure.
19208 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19212 * Extended the ASN.1 parser routines.
19216 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19220 * Added a BN_CTX to the BN library.
19224 * Fixed the weak key values in DES library
19228 * Changed API in EVP library for cipher aliases.
19232 * Added support for RC2/64bit cipher.
19236 * Converted the lhash library to the crypto/mem.c functions.
19240 * Added more recognized ASN.1 object ids.
19244 * Added more RSA padding checks for SSL/TLS.
19248 * Added BIO proxy/filter functionality.
19252 * Added extra_certs to SSL_CTX which can be used
19253 send extra CA certificates to the client in the CA cert chain sending
19254 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19258 * Now Fortezza is denied in the authentication phase because
19259 this is key exchange mechanism is not supported by SSLeay at all.
19263 * Additional PKCS1 checks.
19267 * Support the string "TLSv1" for all TLS v1 ciphers.
19271 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19272 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19276 * Fixed a few memory leaks.
19280 * Fixed various code and comment typos.
19284 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19285 bytes sent in the client random.
19287 *Edward Bishop <ebishop@spyglass.com>*
19291 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19292 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19293 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19294 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19295 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19296 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19297 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19298 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19299 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19300 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19301 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19302 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19303 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19304 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19305 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19306 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19307 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19308 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19309 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19310 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19311 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19312 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19313 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19314 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19315 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19316 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19317 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19318 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19319 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19320 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19321 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19322 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19323 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19324 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19325 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19326 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19327 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19328 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19329 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19330 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19331 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19332 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19333 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19334 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19335 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19336 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19337 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19338 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19339 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19340 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19341 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19342 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19343 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19344 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19345 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19346 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19347 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19348 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19349 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19350 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19351 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19352 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19353 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19354 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19355 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19356 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19357 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19358 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19359 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19360 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19361 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19362 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19363 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19364 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19365 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19366 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19367 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19368 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19369 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19370 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19371 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19372 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19373 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19374 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19375 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19376 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19377 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19378 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19379 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19380 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19381 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19382 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19383 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19384 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19385 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19386 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19387 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19388 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19389 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19390 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19391 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19392 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19393 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19394 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19395 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19396 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19397 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19398 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19399 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19400 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19401 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19402 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19403 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19404 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19405 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19406 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19407 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19408 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19409 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19410 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19411 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19412 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19413 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19414 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19415 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19416 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19417 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19418 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19419 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19420 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19421 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19422 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19423 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19424 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19425 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19426 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19427 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19428 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19429 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19430 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19431 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19432 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19433 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19434 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19435 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19436 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19437 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19438 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19439 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19440 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19441 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19442 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19443 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19444 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19445 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19446 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19447 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19448 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19449 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19450 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19451 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19452 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655